Sea turtle nesting in the Ten Thousand Islands of Florida
Technical Report #56
U.S. Fish and Wildlife Service &
National Park Service
Ahjond S. Garmestani, H. Franklin Percival, Kenneth G. Rice and Kenneth M. Portier
Citation: Garmestani, A.S., H.F. Percival, K.G. Rice and K.M. Portier. 1997. Sea turtle nesting
in the Ten Thousand Islands of Florida. Fla. Coop. Fish and Wildl. Res. Unit, USGS-Biological
Resources Division Tech. Rep. 56. 96 pp.
First and foremost I would like to thank Dr. H. Franklin Percival for providing me
with advice, financial support, folk euphemisms, and the opportunity to work with an
individual as excellent as him. I am also very grateful to Dr. Ken Portier, Dr. Mel
Sunquist and Dr. Alan Bolten for serving as committee members and for offering
valuable advice throughout the course of this research. Special thanks are in order for
SCA/Americorps assistants Kimberley Augenfeld, Kevin LoGiudice, Victor Noerdlinger,
Wayne Taylor, Fred Williams, Jill Witt, and particularly Chuck Lane and Ellen Cheng.
None of the research I conducted would have been possible without their constant
assistance in the field and the laboratory.
Thanks are also in order for Maura Kraus and her staff with the Collier County
Department of Natural Resources and Dr. Todd Hopkins and his staff at Rookery Bay
National Estuarine Research Reserve. Dr. Ray Carthy and Dr. Ken Rice provided
invaluable assistance with research design and analysis. I also would like to extend my
gratitude to Dr. John Davis, Tess Korhnak, Dr. Ken Clark, Dr. Henry Gholz and the
Archie Carr Center for Sea Turtle Research, particularly Dr. Karen Bjomdal, for the use
of lab space and supplies. For their secretarial and administrative support, I am indebted
to Debra Hughes and Barbara Fesler. Others who offered invaluable advice are: Dr.
Clarence Abercrombie, Dave Addison, Craig Allen, Jill Blue, Dr. Mary Collins, Jason
Croop, Dr. Stanley Gehrt, Brendan Godley, Dr. Tim Gross, Dr. Ed Hanlon, Sally
Hopkins-Murphy, Seihun Kong, Dr. JoAnn Mossa, Tom Murphy, Larry Ogren, Dr. Mike
Pelton, Dr. Scott Smith, Phil Wilkinson and John Wooding.
This work would not have been possible without the funding and support of the
Florida Panther/Ten Thousand Islands National Wildlife Refuge (TTINWR), Everglades
National Park (ENP) and the Florida Cooperative Fish and Wildlife Research Unit. In
particular, Ben Nottingham, Jr. (TTINWR), Larry Richardson (TTINWR), Jim
Krakowski (TTINWR), Cindy Garcia (TTINWR), Lennie Jones (TTINWR), Jim
Durrwachter (TTINWR), Mike Mayer (ENP), Daryl Rhodes (ENP), Kathy Clossin
(ENP), Larry Anderson (ENP), Gene Wesloh (ENP), Sid Capo (ENP), Jim Brown (ENP)
and Judy Hayes (ENP) as well as the rest of the staff at ENP and TTINWR, and Sandy
McPherson of the Jacksonville, Florida U.S. Fish and Wildlife Service office. The U.S.
Coast Guard and the National Marine Fisheries Service provided air time in 1996. This
work was funded by the U.S. Fish and Wildlife Service (Cooperative Agreement # 14-16-
0009-1544, Research Work Order No. 119) through the Florida Cooperative Fish and
Wildlife Research Unit (U.S Geological Survey- Biological Resources Division, the
University of Florida, the Florida Game and Freshwater Fish Commission and the
Wildlife Management Institute cooperating).
Loggerhead sea turtles (Caretta caretta) nest in numerous substrate and beach
types within the Ten Thousand Islands (TTI) of southwest Florida. Nesting beach
selection was analyzed on 12 islands within this archipelago. Numerous physical
characteristics were recorded to identify the relatedness of these variables and determine
their importance for nesting beach selection in C. caretta. These variables were chosen
after evaluating the islands, conducting literature searches and soliciting personal
communications. Along transects, data were collected, on the following: height of
canopy, beach width, overall slope (beach slope and slope of offshore approach) and sand
samples analyzed for pH, percentage of water, percentage of organic content, percentage
of carbonate and particle size (8 size classes). Data on ordinal aspect of beaches and
beach length were also recorded and included in the analysis. All of the variables were
analyzed by tree regression, incorporating the nesting data into the analysis. In the TTI,
loggerheads appear to prefer wider beaches (p< 0.001; R2= 0.56) that inherently have less
slope, and secondarily, wider beaches that have low amounts of carbonate (p< 0.001). In
addition, C. caretta favors nest sites within or in close proximity to the supra-littoral
vegetation zone of beaches in the TTI (p< 0.001).
Aerial surveys were flown over the TTI region during June and July of 1996.
Twenty-nine of the possible 34 islands with potential nesting beaches were chosen for the
study. Ground surveys were conducted on 8 islands to provide a correction factor
between ground and aerial counts. All crawls counted on the ground were counted from
the air. This technique resulted in estimates of 906 nests and 997 false crawls in 1996 for
the 29 islands. It appears that sea turtles (Caretta caretta and Chelonia mydas) nest in
relatively high numbers in an area (TTI) not previously known to support significant sea
turtle nesting activity.
Predation by raccoons (Procyon lotor) is the primary factor in nest success for
sea turtles in the TTI archipelago. Eight islands within or adjacent to the Ten Thousand
Islands National Wildlife Refuge (TTINWR) have been surveyed for sea turtle nesting
activity from 1991-96. Raccoon trapping was conducted during the 1992 nesting season
and in 1995-96 from January through April, removing 21, 15 and 2 raccoons,
respectively. Removals in 1992 (21) resulted in continued high predation, indicative of a
high raccoon population and insufficient removals (nests= 42, 95% predation). Although
fewer raccoons were removed prior to the 1995 and 1996 nesting seasons (15, 2), no nest
predation occurred during either year (1995, nests= 41; 1996, nests= 62). These figures
are in contrast to 76-100% nest predation for the previous four years. Eight raccoons
were also trapped, tagged and released on Gullivan Key in 1995 (276 trap-nights, 3.2%
trap success). The following year only 1 raccoon was trapped and removed (114 trap-
nights, 0.9% trap success) resulting in no significant decrease in nest predation (nests=
33, 97% predation). The results from the TTI indicate that raccoon removal may be an
effective management tool for increasing sea turtle nest success on relatively remote
barrier islands, although some raccoons may become sensitized to traps if exposed to
traps prior to removal work. More work is needed to provide conclusive evidence that
sufficient and cost-effective trapping and removal can be conducted on these barrier
island beaches in order to increase sea turtle hatching success.
TABLE OF CONTENTS
1 INTRODUCTION ...............................................................................1
2 EVALUATION OF PHYSICAL PARAMETERS AS..................................... 6
INDICATORS OF NESTING BEACH SELECTION
FOR THE LOGGERHEAD SEA TURTLE (CARETTA CARETTA)
Introduction and Background.............................................................. 6
Homing behavior................................................................... 7
Nesting beach selection............................................................ 9
Nest-site selection................................................................. 11
Objectives...................................... ............................. ... 12
Materials and Methods................................................................... 13
Field Sampling Techniques.................................................... 13
Laboratory analysis............................................................. 17
Statistical analysis............................................................... 18
R esults............................................. ......................................... 18
Nesting beach selection......................................................... 18
Nesting beach selection......................................................... 27
Conclusion .................................................................... 32
3 EVALUATION OF AERIAL SURVEY FOR ESTIMATION.........................34
OF SEA TURTLE NESTING EFFORT IN THE TEN
THOUSAND ISLANDS: APPLICATION OF A
Introduction and Background..........................................................34
Materials and Methods....................................................................35
Study Areas............................................ ........................ 35
Formula ............................................................. ...... .... ..37
R esults...................................................... ..... ......... .................. 38
D discussion ....................................... .......... ................ ... ... ... ........ 4 1
Conclusion ................................................................... ..47
4 MOVEMENT PATTERNS OF THE TEN THOUSAND................................48
ISLANDS RACCOON (PROCYONLOTOR MARINUS)
AND THE EFFECT OF ITS REMOVAL ON SEA
TURTLE HATCH SUCCESS
Introduction and Background..........................................................48
Materials and Methods................................................................... 50
Study Areas....................................................................... 50
1995 Field Season...............................................................52
1996 Field Season...............................................................53
R results .................................................................................... ... 55
1995 Field Season...............................................................55
1996 Field Season................................................................60
Discussion........................................ ....................................... 64
Conclusion................................................. ............... ..... 65
Sea turtles have been the subject of much scientific and public attention due to
their imperiled status and prominent role on beaches of the world. The direct contact
with the public has heightened awareness and concern for sea turtle conservation.
Understanding distribution, nesting and habitat preferences, and sources of mortality is
critical to management of sea turtles and their habitats.
The Ten Thousand Islands (TTI) extend roughly 40 kilometers from Marco Island
to Pavilion Key along Florida's southwest coast (Edwards, 1991). Located
approximately 25 kilometers southeast of Naples, these largely undeveloped islands are
utilized by the public year-round, primarily for fishing (Cheng, 1996). The climate is
subtropical with 110 to 160 centimeters of average annual precipitation, two-thirds of
which occurs between May and October (Cheng, 1996). Average annual temperature is
23 degrees Celsius, with occasional freezing temperatures in the winter and regular mid-
thirties in the summer (Cheng, 1996).
The mangrove ecosystem of the TTI supports a wide range of invertebrates,
fishes, amphibians, reptiles, birds, and mammals. Roughly eighty-six species of fish and
over eighty species of birds have been documented in the area (Cheng, 1996). The
waters serve as a rich nursery ground and an important foraging area for many marine
creatures. Red (Rhizophora mangle) and black (Avicennia germinans) mangroves
dominate most of the islands. Where they exist, beaches are usually narrow and
interspersed with mangroves. Some of the most seaward islands have one or two long
stretches (several hundred meters) of uninterrupted beach. These islands are
characterized by a pioneer zone where vegetation meets the beach. A variety of grasses,
herbaceous plants, and vines are found in this zone, including seashore saltgrass
(Distichlis spicata), sea oats (Uniola paniculata), sea purslane (Sesuvium
portulacastrum), and railroad vine (Ipomoea pescaprae) (Cheng, 1996).
The loggerhead sea turtle (Caretta caretta), is a member of the family
Cheloniidae. It is known to nest in Florida, particularly on the east coast, in the largest
nesting aggregation in the world. Various techniques have been developed to estimate
sea turtle population characteristics. These include a description of techniques for
studying sea turtle population dynamics (Bustard, 1979), a study of population dynamics
of green turtles based on nesting females and hatching (Thompson, 1980) and estimation
of population dynamics for the southeastern United States based on netting of sea turtles
(Henwood, 1987). Chaloupka and Musick (1997) describe the latest techniques for
assessing population dynamics of sea turtles, discussing in detail the numerous variables
that have to be considered when attempting to create a population estimate. Several
researchers have based population estimates on nesting loggerhead females that were
tagged over the course of a 20-year period in Georgia (Crouse et al., 1987; Frazer, 1983;
Frazer, 1984; Richardson et al., 1978; Richardson, 1982).
Numerous studies have been conducted pertaining to sea turtles in Florida, with
some work conducted in southwest Florida. LeBuff & Beatty (1971) and LeBuff (1990)
reported on the nesting status of C. caretta on the Gulf coast of Florida, and LeBuff
(1969) discussed the status of marine turtles at Sanibel and Captiva islands along the
southwest coast of Florida. The studies mentioned above refer to basic descriptions of
the nesting status in southwest Florida in predominantly urban areas. Addison (1994)
commented on the nesting behavior of C. caretta on beaches in the Naples, Florida area,
conducting some work on uninhabited beaches similar to the TTI. LeBuff and Hagan
(1978) assessed the effectiveness of aerial survey as a means of documenting sea turtle
nesting activity on the beaches of the Ft. Myers-Naples, Florida area, with incidental
flights over the TTI included in the study. However, little is known of C. caretta in the
TTI of the west coast of Florida. The TTI have been neglected, with annual ground
surveys conducted by the U.S. Fish and Wildlife Service the only indicator of sea turtle
activity in the TTI. Further down the coast of southwest Florida, Klukas (1967) reported
on the status of marine turtles in Everglades National Park (ENP), while Davis and
Whiting (1977) created a nesting estimate for ENP lands and reported on sources of
mortality for sea turtle nests. Kushlan (1986) repeated the aerial methodology utilized by
Davis and Whiting (1977) in order to create an updated nesting estimate for ENP lands.
The flights over ENP did not cover the TTI in either of the previously mentioned aerial
studies in the TTI. Due to the lack of information concerning sea turtles, it became
important to develop techniques for assessing the nesting status of sea turtles in the TTI.
Included in these techniques are methods for assessing nesting preferences and
distribution of nesting by sea turtles, as well as conservation (i.e. predator control)
Once a sea turtle has navigated to its natal area, it may exhibit nesting
preferences. Only a few researchers directly attempted to clarify reasons for nesting
beach selection in sea turtles. Among these possible factors are an open offshore
approach, sediment type, vegetation, lighting, salinity and area of nesting beach
(Mortimer, 1990, 1995; Johannes and Rimmer, 1984). However, these factors may not
have inter-specific applications, and may differ from study area to study area around the
world. Furthermore, these particular characteristics have been treated individually by
these researchers; they were not subjected to multiple variable analysis.
Miller (1995) reported that sea turtles tend to renest in relatively close proximity
(0-5 km) during subsequent nesting attempts, while a small percentage may utilize more
distant nesting sites. Carr et al. (1978) found that green sea turtles select the same
beaches year after year for nesting, while avoiding other seemingly almost identical
beaches. Melucci et al. (1991) found evidence that hawksbill turtles also exhibit a
significant degree of site-fixity on Antigua. Bjorndal et al. (1985) also found evidence
that the hawksbill turtle exhibits a significant degree of site-specificity at Tortuguero,
although a greater degree of site-fixity existed between seasons rather than within
seasons. Mortimer and Portier (1989) found that green turtles exhibit a high degree of
nest site fidelity at Ascension Island. Tagging data showed that 70 percent of observed
renesting emergences occurred in the same beach cluster as the initial observed
emergence (Mortimer and Portier, 1989). In Chapter 2, I use tree regression to evaluate
numerous physical characteristics of the beaches in the TTI in order to establish
loggerhead sea turtle nesting beach preferences for a particular characteristic or
combination of characteristics. I also discuss nest placement and/or selection by C.
caretta in the TTI.
Aerial surveys have been employed to assess wildlife throughout the United
States and the world. The value of this type of assessment is dependent upon the species
or activity of the species being evaluated. Aerial surveys have been utilized to census sea
turtles and dugongs by several researchers (Marsh and Sinclair, 1989; Bayliss, 1986;
Irvine et al., 1981; Fritts et al., 1983). Marsh and Sinclair (1989) and Bayliss (1986)
reported success in counting dugongs from the air, but reported that due to the diving
behavior of sea turtles, it was difficult to assess their abundance. Fritts et al. (1983)
reported numerous sightings of a variety of species in the TTI region, including large
numbers of loggerheads, and lesser numbers of leatherback, Kemp's Ridley, and green
sea turtles. However, their counts may have coincided with weather and water conditions
that were favorable to spotting sea turtles, as was indicated by Mourao et al. (1994).
Mourao et al. (1994) clearly state that population estimates derived from aerial surveys
are subject to a host of sampling errors. In Chapter 3, I assess the problem of sea turtle
nesting distribution in the TTI. I discuss the possibility of creating a seasonal nesting
estimate for the TTI archipelago based on an aerial survey technique adapted from
Hopkins-Murphy and Murphy (1983).
Raccoons are known to prey on the eggs of numerous species, including:
American crows (Kilham, 1986), waterfowl (Urban, 1970), painted turtles (Christens &
Bider, 1987), wading birds (Frederick & Collopy, 1989), yellow mud turtles
(Christiansen & Gallaway, 1984), coastal burrow-nesting birds (Hartman, 1993) and
diamondback terrapins (Seigel, 1979). The impacts of raccoons on sea turtles are well
known throughout the United States, particularly in areas where continuous studies have
been conducted. The Georgia Department of Natural Resources assessed the impacts of
raccoons, as well as feral hogs on sea turtle nests on several barrier islands for six
seasons. Raccoons have been shown to have definitive impacts on the success of sea
turtle nesting over the six-year period (Harris and Maley, 1990; Maley and Harris, 1991,
1992; Maley and Murphy, 1993, 1994; Maley, 1995). Raccoons have been determined to
be a major factor in the success of sea turtle nesting in the TTI of southwest Florida (L.
Richardson, pers. comm.). The TTI are inhabited by the Ten Thousand Islands raccoon
(Procyon lotor marinus), a subspecies of P. lotor which was first identified by Nelson
(1930). P. 1. marinus is characterized by a smaller stature than the mainland subspecies,
as well as cranial differences (Nelson, 1930; Bigler et. al., 1977). Due to the high
incidence of depredation of sea turtle nests by raccoons (ranging from 59-85% for 1991-
1994) (Garmestani, 1995), a better understanding of raccoon re-population trends,
movement patterns and effects of removal on sea turtle hatch success in the Ten
Thousand Islands National Wildlife Refuge was necessary to create a more
comprehensive sea turtle management strategy (Shabica et al., 1978). I discuss this
research in Chapter 4 and offer management suggestions for raccoons in regard to sea
Finally, in Chapter 5, I suggest management options for the TTI with respect to
sea turtle conservation.
EVALUATION OF PHYSICAL PARAMETERS AS INDICATORS OF NESTING
BEACH SELECTION FOR THE LOGGERHEAD SEA TURTLE
Introduction and Background
The Ten Thousand Islands (TTI) have been neglected when assessing sea turtle
populations in southwest Florida. No accurate estimates have been made for the region
because surveys conducted in the archipelago have been incomplete. Only the area
encompassed by the Ten Thousand Islands National Wildlife Refuge (TTINWR) has
been effectively monitored. TTINWR accounts for approximately 20% of the possible
turtle nesting beaches in the archipelago; the remainder is already under the ownership of
Everglades National Park (ENP) and the Florida Department of Environmental
The TTI are characterized by undeveloped mangrove islands with low human
activity, predominantly fishermen and campers (fall, winter and spring). The islands are
not subject to the kind of nesting disruption due to human usage that affects distribution
of loggerhead nests in other research areas (e.g., North Carolina; Fangman and
Rittmaster, 1994). Fangman and Rittmaster (1994) found that loggerhead sea turtles
tended to abort nesting attempts with the increased presence of humans (e.g., camping,
night-walking, parties). With the continued protection of these islands by the federal
government (National Park Service, U.S. Fish and Wildlife Service) and the Department
of Environmental Protection (Rookery Bay National Estuarine Research Reserve), the
relatively unspoiled turtle nesting beaches offered a unique opportunity for study.
Because of the potential significance of the TTI for sea turtle nesting, further exploration
of these islands and accurate predictions on sea turtle nesting was necessary to determine
the status of turtle populations in this area.
Information gathered for eight monitored islands in the refuge indicated that the
total nesting effort for the Ten Thousand Islands could be significant (Table 2-1).
Table 2-1. Total TTINWR turtle survey data (1991-1995)
Year # of crawls # of known # of depredated % of
nests nests depredated
1991 194 157 128 82
1992 255 139 118 85
1993 189 93 61 66
1994 232 157 92 59
1995 364 206 85* 41*
*- includes raccoon removal island (Panther Key)
Sea turtles exhibit three basic behaviors when returning to natal areas to nest:
homing behavior, nesting beach selection and nest-site selection. Homing behavior is
discussed as a means of inter-connecting these behaviors, although it was not examined
in this study. Nesting preferences are discussed in detail, with nesting beach selection the
focus of this research.
Sea turtles are known to exhibit varying degrees of site-specificity to their natal
areas. Grassman et al. (1984) found evidence that sea turtles detect differences in water
samples when exposed to different samples in a laboratory setting. In laboratory
experiments, the researchers found that sea turtles preferred to spend time in water
solutions in which they had been artificially imprinted. This research suggests that sea
turtles may have the ability to identify their natal areas by imprinting on olfactory cues
specific to that particular area. Lohmann (1992) commented that tagging studies at
Ascension Island have shown that green turtle females remain faithful to their nesting
site; no turtle seen at Ascension has ever been found nesting elsewhere. Loggerhead
turtles also exhibit some nest site-fixity (Lund, 1986), as evidenced by the 49% return of
tagged female turtles over ten years at Cumberland and Little Cumberland Island,
Georgia (Richardson et al., 1978). The methods by which sea turtles orient and return to
natal areas has been the subject of much debate and research. Lohmann (1992) has
suggested that cues utilized by other migratory animals such as the position of the sun or
stars, polarized light, odors, wind direction, infrasound and/or the earth's geomagnetic
field may play a role in navigation for sea turtles. Lohmann (1992) does not rule out the
possible reliance upon chemosensory factors, as suggested by Grassman et al. (1984), or
other as of yet unidentified cues, but hypothesizes that they are not a primary factor in
Lohmann (1992) suggested that adult sea turtles appear to fix their position with
respect to their destination. Furthermore, he discounted the use of stars as an orientation
cue based on anatomical studies of sea turtle eyes that revealed them to be extremely
myopic when their heads are above water (Ehrenfeld and Koch, 1967; Koch et al., 1969).
He further suggested that adults may have a "map sense", that allows them to orient
according to the magnetic field line inclination, or by detecting pathways along the ocean
floor created by magnetic maxima and minima. As with hatchlings, wave direction may
be a possible cue, due to the independence of swell direction from local weather patterns,
for adult turtles migrating back to nesting areas.
The idea of "social facilitation", suggested by Owens et al. (1982), was refuted by
conducting mtDNA analyses on three different sea turtle species (Chelonia mydas,
Caretta caretta and Eretmochelys imbricata) (Allard et al. 1994; Bowen et al., 1992;
Bowen et al., 1993; Broderick et al., 1994; Meylan et al., 1990). Bowen (1995) stated
that the mtDNA evidence suggests that nesting populations are distinct demographic
units, and depletion of nesting aggregates will not be compensated by recruitment from
other populations. What this evidence suggested is that sea turtles are fixed upon a
particular natal area and return to that area utilizing a navigation mechanism that is not
yet understood. These aforementioned studies dealt with some of the possible
mechanisms sea turtles may use to return to their natal area, in this case the TTI and/or
southwest Florida, but further exploration was necessary to indicate preferences for
Nesting beach selection
Shoop et al. (1985) made the observation that a high-density loggerhead sea turtle
nesting beach was abandoned over the course of a five-year period due to, in their
estimation, a mud flat that developed on the ocean side of the island. This sort of drastic
change in a suitable nesting habitat does not necessarily affect the homing behavior of sea
turtles to a particular area, but it would drastically affect nesting beach selection.
Therefore, it appears that certain chemical and/or physical factors influence loggerhead
sea turtle nesting beach selection. Mortimer (1995) has conducted the most
comprehensive research dealing with nesting beach preferences in sea turtles. She
conducted research on green turtles at Ascension Island, analyzing numerous components
of the island in an attempt to determine nesting beach preferences for C. mydas.
Mortimer (1995) commented that there are some basic requirements indicative of a good
nesting beach. Among these characteristics are: easy accessibility from the sea, a beach
platform that is high enough to not be inundated by high tides or the water table, and
beach sand that facilitates gas diffusion, but moist enough and fine enough to prevent
excess slippage during nest construction (Mortimer, 1995). Taking these factors into
account, Mortimer (1995) evaluated beach length, rock cluttering the foreshore, rock
obstructing the offshore approach and artificial light visible on the shore as possible
indicators of nesting beach selection.
Sea turtle nesting preferences for C. caretta in the TTI were assessed by
evaluating each nesting beach based on a subset of physical characteristics of the islands.
These characteristics included: height of canopy (Bustard, 1972); beach width, overall
slope (beach slope and slope of offshore approach) (Mortimer, 1995; Shoop et al., 1985)
and sand samples analyzed for pH, percentage of water, percentage of organic content,
percentage of carbonate and particle size (8 size classes) (Mortimer, 1995; Mortimer,
1990; Stancyk and Ross, 1978; Hughes, 1974). Data on ordinal aspect of beaches and
beach length (Mortimer, 1995) were also recorded and included in the analysis. These
factors were based upon characteristics of the islands, past research and personal
communications (R. Carthy, A. Bolten, K. Rice and H. Percival). Although green sea
turtles were known to nest in the TTI, they were not included in this study.
Numerous studies have been conducted on nest-site selection in sea turtles,
addressing various factors as potential indicators of nest-site selection. Johannes and
Rimmer (1984) found that nest sites for the green turtle were on beaches that had lower
salinity at nesting depth, lower salt content at the surface of the sand and were sheltered
from prevailing winds. Stoneburner and Richardson (1981) found that sea turtles cross a
temperature gradient after they have landed on a nesting beach. A nesting female will lay
her eggs once she has passed over a zone of the beach where a rapid rise in temperature is
encountered. However, Hays et al. (1995) were critical of the latter study, indicating that
the sand temperatures Stoneburner and Richardson (1981) recorded were from disturbed
sand at C. caretta nest sites. Hays et al. (1995) suggested that the abrupt temperature
increase at nest sites may be a result of deeper, warmer sand being brought to the surface.
They were also critical of the lack of explanation for this abrupt temperature increase (i.e.
no correlation with any of the features of the beach topography). Hays et al. (1995)
compared nest-sites for loggerhead turtles in southwest Florida, and green turtles on
Ascension island, and found that loggerheads clump their nesting attempts close to
vegetation backing nesting beaches, while green turtles appear to cue on uneven beach
topography when choosing nesting sites. Stancyk and Ross (1978) found no correlation
between sand characteristics of the beaches at Ascension Island and nesting frequency of
green turtles. Mortimer (1990) also conducted a study at Ascension Island and found no
preference for a particular sand type based on clutch survival, and concluded that factors
other than the physiognomy of the sand on nesting beaches may be more important in
nest-site selection for the green turtle. Bjorndal and Bolten (1992) concluded that annual
factors may have a greater effect on nest placement than individual nesting patterns for
green turtles at Tortuguero. This lack of consistent nest distribution may result from
environmental uncertainty and low predictability of nest success (Bjorndal and Bolten,
1992). Mrosovsky (1983) was unable to explain the poor nest-site selection of
leatherback turtles in Guianas and Malaysia. In addition, Camhi (1993) conducting
research at Cumberland Island, Georgia found that loggerhead turtles did not show strong
site preferences, even though certain areas of the beach were more likely to produce
hatchlings. Whitmore and Dutton (1985) found that leatherbacks nested more in open,
sandy areas, while green turtles nested in predominantly vegetated areas. Daud and
deSilva (1987), found that leatherbacks preferred to nest in fine sand on steep sloped
beaches, while green turtles preferred to nest in coarser sand on broad, flat beaches.
Nest-site selection was also evaluated for C. caretta utilizing nesting data from
the 1996 field season. No attempt was made to evaluate the possible cues indicating nest-
The primary objective of this research was to establish if the loggerhead sea turtle
exhibits preferences for a particular physical characteristic, or combination of
characteristics when selecting nesting beaches. The secondary objective of this research
was to document the overall distribution of loggerhead sea turtle nests and determine if
there was a significant trend of nest placement for the TTI.
Materials and Methods
The TTI extend roughly 40 kilometers from Marco Island to Pavilion Key along
Florida's southwest coast (Edwards, 1991). Located approximately 25 kilometers
southeast of Naples, Florida, these largely undeveloped islands are utilized by the public
year-round, primarily for fishing (Cheng, 1996). The climate is subtropical with 110 to
160 centimeters of average annual precipitation, two-thirds of which occurs between May
and October (Cheng, 1996). Average annual temperature is 23 degrees Celsius, with
occasional freezing temperatures in the winter and regular mid-thirties in the summer
The study area included 12, low-relief mangrove islands, further broken down
into 35 beaches based on the ordinal aspect of the individual beaches. Islands in the
study area directly face the Gulf of Mexico and were known to support sea turtle nesting.
These islands included: Brush, B, Turtle, Gullivan, White Horse, Hog, Panther and
Round Keys, located in the Cape Romano Aquatic Preserve and Ten Thousand Islands
National Wildlife Refuge, and Kingston, Indian, Picnic, and Tiger Keys in Everglades
National Park. These islands are located to the southeast of Marco Island, and are part of
the western delta of Florida's Everglades. Beaches are made up of differing substrate
types that are usually backed by a variety of supralittoral vegetation (Table 2-2). The 35
beaches examined totaled 7.404 kilometers in total length.
Field Sampling Techniques
Twelve (12) islands were selected for study based on logistical constraints and
supporting agencies' (TTINWR and ENP) preferences. This included the 8 islands
covered in the five-year monitoring period for TTINWR (Edwards, 1991) and 4
additional islands within ENP. Although these islands do not represent a random sample
of possible islands in the TTI, they are representative (Garmestani, pers. observe ) of the
range for each of the variables selected for analysis. The 12 islands are a sample of the
19 northernmost keys, with 10 other keys in the TTI also included in the generalization of
the results of this research.
Sea turtle surveys were conducted from May 13, 1996 until August 15, 1996.
Surveys were conducted twice a week for each island. This strategy was developed by
the U.S. Fish and Wildlife Service, due to the remoteness of the TTI and lack of
personnel. A total of 252 loggerhead nests was used in the analysis for nesting beach
Table 2-2. List of plant species sampled within quadrats at
Wild potato vine
e Avicennia germinans
n Caesalpinia bonduc
* Casuarina equisetifolia
i Chrysobalanus icaco
rass Distichlis spicata
S Ipomoea pandurata
e Languncularia racemosa
a U Ludwigia palustris
zh Mentzelia floridana
et Yucca aloifolia
selection, while a subset of 236 nests was used in the analysis for nest-site selection.
Crawls were described and data were collected from each nest site. Included in the data
set were species of turtle, vegetation (if any), distance of nest from mean high tide line
(MHTL), distance of nest from the vegetation (calculated as a percentage of the beach
width for each nest), distance of nest from the nearest transect and status of nest (e.g.,
Island characteristics were measured in April and May of 1996, prior to sea turtle
nesting season. Each of the 12 islands was assessed for a subset of physical
characteristics that might influence loggerhead nesting beach selection along 82 transects
that were placed at 100m intervals along nesting beaches. Each transect was run
perpendicular from the beach/vegetation transition (i.e. starting point) to a point 100m
offshore using a compass to maintain the perpendicular vector of the transect, and a
survey tape to record distance. The beach/vegetation transition was defined by the first
occurrence of vegetation on the open beach and marked with a stake along the transects.
Data were collected on the following characteristics:
1) Beach width Beach width was defined as the distance from the MHTL to the
beach/vegetation transition, and measured in meters along the transects using a distance
2) Slope Slope was assessed along the transects from each of the individual
stakes. Sample point distances were taken in meters with a survey tape. Depth
measurements were made in meters using a laser transit and survey pole, or in the case of
deeper areas, a Lowrance sonar (X-16 paper graph unit; Lowrance Electronics, Inc.,
Tulsa, OK, USA). Measurements were taken in increments based upon the rate of
change of the slope. Least squares technique was used to estimate slope based on linear,
quadratic or cubic models.
3) Height of canopy Canopy height was defined as the height of the vegetation
taken while standing at MHTL, and observing the canopy with a clinometer. This
characteristic was assessed along each transect using the clinometer and a survey tape,
and measured in meters.
4) Nesting substrate Nesting substrate samples were taken at MHTL along each
transect (i.e. shell content, particle size, organic content, water content and pH levels).
Samples were taken at MHTL, because after landing a sea turtle may exhibit a preference
for a particular nesting beach based on factors in the sand with which she comes in
contact as she initially leaves the surf for the open beach. The nesting substrate samples
were collected by digging vertically and taking a 300 gm sample, placing the sample in a
ziplock bag and storing the sample on ice, to preclude chemical changes in the samples.
The samples were maintained on ice as they were transported to the field headquarters in
Everglades City and refrigerated until they were taken to Gainesville on ice. After
arriving in Gainesville, the samples were refrigerated until they were analyzed in the
5) Ordinal aspect Ordinal aspect is defined by the angle of the vector running
parallel to a nesting area. This feature was assessed with US Geological Survey, Collier
County and NOAA maps, and compass measurements. The 12 islands were divided into
35 "beaches" based on the ordinal aspect of their nesting areas.
6) Beach length Beach length was defined as the horizontal cross-section
distance of each beach, and was measured in meters using a distance measuring wheel
7) Vegetation Nesting area vegetation was classified according to common
name and scientific name at each transect starting point in a 1m2 quadrat.
Sand samples were assessed for pH, particle size, water content, organic content
and carbonate content. For pH analysis, two 5g sub-samples were taken. Since sea
turtles would come into contact with wet sand, one was assessed directly, while the other
was processed in a drying oven at 105 degrees Celsius for 12-16 hours as per standard
soil science technique. Each sample was mixed with de-ionized and distilled water in a
1:2 mixture, and allowed to sit for one hour. Then the mixture was stirred and a pH
reading was taken and recorded.
For particle size, a 50cc. sub-sample was air-dried for 72 hours and was shaken
through 8 sieves (< 0.246mm, 0.246, 0.351, 0.495, 0.701, 0.991, 1.397, 1.981mm). Each
size class was weighed and the results recorded. For water, organic and carbonate
content, one 25cc. sub-sample was taken and subjected to analysis outlined in Dean
(1974) and Schulte (1988). The 25 cc. sub-sample was dried in a drying oven at 105
degrees Celsius for 12-16 hours, allowed to cool to room temperature in a desiccator and
then weighed to determine water content. The dried sub-sample was tested for organic
content by heating in a muffle furnace for 24 hours at 360 degrees Celsius and weighing
after cooling to room temperature in a desiccator. The resulting sample was processed by
heating in a muffle furnace at 850 degrees Celsius for 4 hours, and weighing after it had
cooled to room temperature in a desiccator, to determine carbonate content.
Sand samples for organic and carbonate content were assessed in triplicate in
order to validate consistent laboratory technique, as well as account for variability among
sub-samples. The average values obtained from the laboratory analysis were utilized for
The multi-dimensional data which resulted from the field measurements and
laboratory analysis were plotted in their first two principal components to detect obvious
clustering of transects. The relationship between beach utilization for nesting and beach
characteristics was examined using a robust tree regression technique (Portier and
Anderson, 1995). The final tree regression model was used to indicate which
characteristics should be considered important in predicting beach use by nesting turtles.
A binomial test was used to compare nest counts on the beach to nest counts in the
vegetation. A Chi-square goodness-of-fit test was used to determine if nesting occurred
uniformly across the width of the beach.
Computations associated with the formal statistical tests and computations of
beach slope were performed in SAS (SAS Institute, Inc, 1989) on a personal computer.
The tree regression and principal components analysis computations were performed
using S-PLUS (Statistical Sciences, 1993; Venables and Ripley, 1994) on a UNIX
Nesting beach selection
The principal components plots did not demonstrate significant clustering within
the data. A more formal cluster analysis using the average linkage method also failed to
The final regression tree (R2= 0.56; Figure 2-1) identified beach width (p< 0.001)
as the primary factor affecting nesting beach selection for C. caretta. Beaches that are
wider (>8.5m; n= 15) on average have much higher nesting than narrower beaches
(<8.5m; n= 20). Beach width was negatively correlated (r= -0.6) with beach slope, wider
beaches having inherently less slope. For wider beaches (> 8.5m) the regression tree
indicated a further split using percentage of beach carbonate (p< 0.001). On wide
beaches loggerheads preferred less than 9% carbonate (n= 10) (i.e. function of the
shelliness of a beach), over beaches that had over 9% carbonate (n= 5). Wider beaches
having low carbonate levels will be expected to have the highest nesting. The percentage
of carbonate (i.e. shelliness) was also positively correlated (r= 0.88) with the three largest
particle size classes (0.991mm, 1.397mm, 1.981mm). Narrow beaches (n= 20) (< 8.5m
beach width) had 153 crawls, of which 60 (39%) resulted in nests. Wide beaches (n= 5)
(> 8.5m beach width) with high levels of carbonate (> 9% carbonate), received 50 crawls,
l5 a5 o
J 0 v
a o-g i g
^~ v- ^
of which 21 (42%) resulted in nests. Finally, beaches that were wide (> 8.5m) and had
less than 9% carbonate (n= 10), had a total of 292 crawls, of which 171 (59%) resulted in
nests. The number of nests on beaches increased dramatically in conjunction with
increased beach width (Figure 2-2). The nesting to non-nesting emergence ratio was
decidedly higher on beaches possessing the favored characteristics (Figure 2-3). All of
the beach data are summarized in Table 2-3.
The other variables included in the analysis were not as highly correlated to
loggerhead sea turtle nesting beach selection as were the primary and secondary
components. Although these variables may be of tertiary importance in nesting beach
selection, there influence will be much less than the two characteristics identified in the
regression tree. With a larger data set it might be possible to determine which of these
variables further discriminate nesting beach selection.
After running the duplicate analyses on the sand sub-samples, it was determined
that there was greater than 2% variability between several of the samples. Therefore, it
was necessary to run a third sample set in order to get an average of the three samples and
alleviate concerns over variability between the samples. Overall, the triplicate samples
had a 9.2% coefficient of variation.
There were 236 loggerhead nests used in the analysis, 111 of which were located
in the vegetation and 125 on the open beach (Figure 2-4). No significant preference
between open beach and vegetation was detected for C. caretta nest placement (p= 0.82,
binomial test). However, location of nests was not uniformly distributed in relation to the
vegetation and vegetation lines on beaches (p< 0.001, Chi-square test; Figure 2-5). Of
these nests (n= 236), 127 were depredated, 106 hatched and 3 were inundated. For the
depredated nests (n= 127), there was no significant difference in depredation rates
detected for nests in the vegetation (n= 59) or nests on the open beach (n= 58) (p= 0.14).
Beach width (m)
Figure 2-2. Number of nests on TTI beaches in relation to beach width
6 0o 5
A M A*-
(seoue~jweu -0 equlnu)
^0 rt~~ 03 -
:~ C. Il
e~euoqje~ o 10e~elueojed pue Lfp|M ipeeg
Table 2-3. Index of physical components for Ten Thousand Islands
Beach slope 82
Offshore slope 82
Beach width 82
Height of canopy 82
pH (wet) 82
pH (dry) 82
% H20 82
% Organic material 82
% Carbonate 82
% Sand 1 (1.981mm) 82
% Sand 2 (1.397mm) 82
% Sand 3 (0.991mm) 82
% Sand 4 (0.701 mm) 82
% Sand 5 (0.495mm) 82
% Sand 6 (0.351 mm) 82
% Sand 7 (0.246mm) 82
%/ Sand 8 (<0.246mm) 82
Beach length 35
Ordinal aspect 35
N Mean S.E.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Distance from vegetation line (m)
Figure 2-4. Distance (m) of loggerhead nests to supra-littoral vegetation line
) Io w I N
sisa JO T-Wfl
'. ; / .' **.0 ?)
.__ ^. ._ -rf,,a.joe- (
^ *-a ..*' ^ -- .^ ^V I. (
There was no significant difference in the hatch rate for the hatched nests (n=
106) whether they were in the vegetation (n= 51) or on the open beach (n= 55) (p= 0.39).
It is now widely accepted that sea turtles have the ability to home to particular
areas of the world's seas for feeding, mating and nesting. Lohmann (1992) has described
mechanisms of navigation utilized by hatchling sea turtles, and has hypothesized about
the possible mechanisms that adults use to home to their natal areas for nesting. The
distinctiveness of sea turtle populations as demographic units indicates that sea turtles are
programmed to home to natal areas (Bowen, 1995). Meylan et al. (1990) concluded that
social facilitation to non-natal sites was rare due to evidence of mtDNA sequences for
turtles sampled. Allard et al. (1994) supported the results found by Meylan et al. (1990)
by conducting a study on green turtles at Tortuguero, Costa Rica and the southeast coast
of Florida. They found evidence, in the form of mtDNA sequences, that the Costa Rican
and Florida populations were significantly different (Allard et al., 1994).
On a finer scale, Murphy and Hopkins-Murphy (1990) conducted research on the
homing behavior of gravid loggerhead turtles in South Carolina. Of the 27 radio-tagged
turtles they collected data from, 23 (85%) showed evidence of returning to their home or
primary beach (Murphy and Hopkins-Murphy, 1990). They found that there is a
plasticity to site selection which allowed for the acceptance of suitable alternate nesting
beaches for one or more nesting events for several of the turtles (Murphy and Hopkins-
Murphy, 1990). They hypothesized that loggerhead sea turtles may select beaches
according to offshore topography and/or the suitability of the beaches for nesting at the
translocation site (Murphy and Hopkins-Murphy, 1990). They suggest that loggerhead
sea turtles may be exhibiting preferences when selecting nesting beaches. Carr and Carr
(1972) reported that green turtles showed strong site fixity, although some re-nesting
attempts stray up to 7 km from previous nesting attempts. They suggest that colony
expansion might occur in this manner (Carr and Carr, 1972). Another possible
explanation for this phenomenon is that green turtles have a preference for particular
beach characteristics. After homing to natal areas, sea turtles may have several beach
types from which to choose, and select nesting beaches according to criteria specific to
their natal area.
Nesting beach selection
Mortimer (1995) reported that the heaviest nesting by green turtles on Ascension
Island occurred on beaches that were unlighted, with open offshore approaches and
foreshores relatively free of rock clutter. In addition, most nesting occurred on stretches
of beach where the offshore approach was deepest (Mortimer, 1995). For the beaches
that met these criteria, the correlation between beach length and the estimated number of
clutches laid per season was highly significant (Mortimer, 1995). Mortimer (1995) found
no relationship between beach vegetation or beach sand characteristics and density of
nesting. Johannes and Rimmer (1984) reported that no relationship has been found
between nesting beach selection and pH, calcium carbonate content, water content,
organic content and particle size. However, they found beaches favored by the green
turtle in Australia had lower salt content and less exposure to prevailing winds (Johannes
and Rimmer, 1984). Although Johannes and Rimmer (1984) found a positive
relationship between these characteristics and green turtle nest density, there is no
evolutionary explanation for the observed phenomenon. While sea turtles could possibly
detect elevated levels of salinity upon immediate landing on a nesting beach and choose
another nesting beach, there is no clear relationship between elevated levels of salinity
and sea turtle nesting beach selection. There is no literature referring to the potential
impacts of increased salinity on egg development in sea turtles. One of the only
references in the literature to salinity and its effects on development, deals with its
impacts on egg development in Baltic cod (Westin and Nissling, 1991). Therefore, the
connection between salinity and nesting beach selection is questionable. Furthermore,
the connection between prevailing winds and nesting beach selection is not clear. This
point is reinforced by the fact that prevailing winds in the study area frequently ceased at
night, making them a non-factor in nesting beach selection (Johannes and Rimmer,
The tree regression approach was used in the analysis because it is very flexible,
allowing the data to specify complex interactions where justified, and very simple to
interpret. This approach is also less concerned with underlying distributional
assumptions than are traditional linear regression approaches. Tree regression models
expected nest counts as a binary decision process, which is a result of comprehensive
examinations of all levels of each characteristic in a hierarchical model-building process.
The process involves constructing a large binary tree, then pruning the tree back to a size
which is consistent among sub-samples of the complete data set.
For the loggerhead sea turtle, a positive relationship was found between beach
width, which is highly correlated with slope (i.e. wider beach=decreased slope), and the
number of nests (n= 192; 76%) laid on beaches. On these wider beaches, loggerheads
preferred those with a lower shell content, which is highly correlated with the three
largest particle size classes analyzed. The fact that the number of non-nesting
emergences increased on beaches that do not possess the favored criteria only lends
strength to the tree model.
Loggerheads may prefer wider beaches because the loss of nests from tidal
inundation or erosion decreases with distance from the surf (Fowler, 1979). Choosing
beaches that are relatively wide may be the only alternative for a sea turtle attempting to
produce a successful clutch in a region like the TTI, where beaches are narrow and
subject to wide variation in tidal fluctuation from season to season and year to year.
Slope may simply be a construct of beach width in the area, or C. caretta may be cueing
on the slope of the beach. This makes sense, because the gradient of the beach
determines the distance a turtle must crawl overland in order to reach a nest site
(Pritchard, 1971; Schulz, 1975). Loggerheads may prefer more gently sloping beaches
because they afford the quickest and easiest approach and means of escape after laying a
nest. Bjorndal and Bolten (1992) have suggested that in long-lived species that cannot
accurately assess the nesting environment, it may be better to select to nest on the basis of
the survivorship of the adult female rather than on the survivorship of the clutch.
C. caretta may choose to nest on wide beaches that are relatively free of large
shell fragments because the shells may make it more difficult for the turtle to create an
adequate nest cavity. Digging in shells is difficult, even for a creature as large and
powerful as a loggerhead sea turtle. Furthermore, nesting cavities made up of shells will
be more prone to collapse as the turtle excavates the chamber. Considering that the
quality of beach sand can so strongly influence reproductive success, one would expect
turtles to use sand textures as a criteria in selection (Mortimer, 1990). Data gathered by
Mortimer (1995) at Ascension Island suggest otherwise. There was no correlation
between the average percent hatchling emergence at Ascension Island beaches and
nesting density (Mortimer, 1995). It appears that the choice of nesting beach is primarily
determined by the ease with which a nest site can be reached and excavated, with little
choice exhibited for more favorable reproductive substrates.
No significant relationship was found between pH, percentage of water,
percentage of organic content, or the 5 smallest classes of the 8 particle size classes and
the number of nests laid on beaches in the TTI. From these data, it appears that C.
caretta does not select nesting beaches according to the height of the canopy immediately
adjacent to nesting beaches, the beach length, the slope of the offshore approach, or the
ordinal aspect of the beaches. Although some of these factors could potentially affect the
reproductive success of sea turtles, none affected the nesting beach selectivity of C.
Nest-site selection has been the subject of much sea turtle research (Bjomdal and
Bolten, 1992; Camhi, 1993; Carr and Carr, 1972; Daud and deSilva, 1987; Miller, 1995;
Mrosovsky, 1983; Whitmore and Dutton, 1985). Numerous theories have been suggested
as possible explanations for this critical process in sea turtle ecology, with multiple
variables assessed as potential indicators of nest-site selection. Among these are salinity
(Johannes and Rimmer, 1984), temperature (Stoneburner and Richardson, 1981), particle
size, pH, water content, organic content, shell content, substrate conductivity (Mortimer,
1990; Stancyk and Ross, 1978) and vegetation (Hays and Speakman, 1993). A
connection has been made between temperature, salinity and nest-site selection (Johannes
and Rimmer, 1984; Stoneburner and Richardson, 1981), and some researchers have
suggested the vegetation zone backing beaches as a potential cue for sea turtles to select
nest sites (Hays and Speakman, 1993). The importance of this process on the
reproductive success of the turtle is critical, however, sea turtles often choose nest sites
that do not maximize their reproductive output, and therefore their fitness. When it is
apparent that certain beach types drain better, provide better gas and water exchange and
are more favorable for maintaining a nest cavity (Pollock and Hummon, 1971), it is hard
to imagine that sea turtles are not programmed to detect these obvious advantages in
selecting certain nest sites over others (Mortimer, 1995). Bolten and Bjorndal (1992)
have suggested that a female may have a limited ability to assess the current nesting
environment when selecting a nest site, and unpredictable changes in the nesting
environment over a 60-day incubation period. They believe this environmental
uncertainty and the inability of female turtles to assess the nesting habitat in terms of
ultimate hatching success may be responsible for the lack of any consistent patterns in
nest distribution for green turtles at Tortuguero (Bjorndal and Bolten, 1992). Hays and
Speakman (1993) found that loggerheads tended to lay nests away from the sea, and close
to, but not beyond the vegetation line. The increase in the distance turtles laid from the
sea when the vegetation line was further from the water suggests that the vegetation may
have constrained the length of the turtle's inland crawl (Hays and Speakman, 1993).
Some researchers have suggested that loggerhead sea turtles simply crawl a random
distance above the most recent MHTL prior to digging a nest site (Hays et al., 1995).
They constructed a random-crawl-distance model that closely reproduced the observed
spatial pattern of nests (Hays et al., 1995). Hays et al. (1995) stated that the distribution
of nests on Sanibel and Captiva islands in southwest Florida can be explained by the
random-crawl-distance model without implicating micro-habitat cues.
Although it has been suggested that supra-littoral vegetation backing nesting
beaches may act as a cue for nesting sea turtles (Hays and Speakman, 1993),
documentation of nest placement by C. caretta in the TTI indicates that no such
relationship exists. Of the nests analyzed for 1996 (n= 236), 111 were laid within the
supra-littoral vegetation backing nesting beaches, while 125 were laid on the open beach
at varying lengths from the vegetation line. The majority of the 125 nests laid on the
open beach were clumped in close proximity to the vegetation line. Furthermore, nests
laid in the vegetation suffered nearly equal depredation rates when compared to nests laid
on the open beach, and therefore no appreciable difference in hatch rates. Since no
micro-habitat cues were analyzed for this research, it is difficult to explain why C. caretta
nests in this manner in the TTI. The most obvious explanation for this phenomenon is
that loggerheads may select nest sites that will lessen the chance of tidal inundation or
inundation of nests by the water table (Fowler, 1979). The possibility exists that they
simply crawl a random distance beyond the last MHTL and lay their clutch, as has been
suggested by Hays et al. (1995). No attempt was made to verify the theory proposed by
Hays et al. (1995); the theory only exists as one possible explanation for nest placement
by the loggerhead sea turtle in the TTI.
The TTI is a unique ecosystem, and results from this study are unlikely to be
directly applicable to other loggerhead nesting areas. The same is true of the green turtle
work conducted by Mortimer (1995) at Ascension Island. Therefore, the conclusions
reached about green turtles and loggerhead turtles may be specific to the environments
evaluated. Without testing these methodologies on turtles in other parts of the world,
there is no way of validating the results. Therefore, the only way of commenting on
loggerhead nesting beach selection in other parts of the world is to replicate this
methodology in another part of the world where C. caretta nests.
Loggerhead sea turtles in the TTI appear to have chosen nesting beaches
according to the width of the beach, preferring relatively wide beaches. Secondarily, C.
caretta preferred wider beaches according to the shelliness (i.e. carbonate content) of the
beach. C. caretta also selected nest-sites in the TTI that are in or close to supra-littoral
From a management perspective, this means that biologists/conservationists in the
TTI could concentrate their efforts for loggerhead sea turtle conservation on beaches that
meet the criteria for favorable nesting beach selection. Beach width measurements could
be taken, and sand from the beaches could be shaken through a series of sieves in order to
assess the quality of a particular beach. Beach width could be assessed quickly and
easily; beaches that are on average greater than 8.5 meters in width should be selected.
After these wide beaches were selected, the resource manager could judge the amount of
shells (i.e. carbonate content) that remains in the three largest sand classes (0.991mm,
1.397mm, 1.981mm) after the sample has been shaken through the sieves. If there is a
large proportion of shells (> 9%) in the sample, then the manager has the ability to make
decisions about allocation of time, money and personnel to particular islands and/or
beaches, based on the quality of the particular nesting area for C. caretta.
This management option is presented as a possibility for application in the TTI.
However, at least one more year of data would need to be collected in order to determine
if the nesting beach selection observed in 1996 represented a trend in the behavior of C.
caretta. If the replicate study re-affirms the findings of this study, then the suggested
management scheme could be implemented on a trial basis.
EVALUATION OF AERIAL SURVEY FOR ESTIMATION OF SEA TURTLE
NESTING EFFORT FOR THE TEN THOUSAND ISLANDS: APPLICATION OF A
Introduction and Background
The remoteness of the Ten Thousand Islands (TTI) provided a unique opportunity
to conduct biological research in a relatively undisturbed setting. However, this desirable
characteristic (i.e. remoteness) of the TTI limits access to the area, requiring large
expenditures of personnel, time and expenses to effectively monitor and protect sea turtle
nests. Numerous researchers have used aerial surveys to assess large areas of sea turtle
nesting beaches. These include an assessment of sea turtle nesting in the southeastern
United States (Shoop et al., 1985), North Carolina (Crouse, 1984), South Carolina
(Stancyk et al., 1979), Baja California (Fritts et al., 1982), Panama (Meylan and Meylan,
1985) and a survey of Pacific Mexico coupled with a new estimate of the world's
leatherback (Dermochelys coriacea) turtle population (Pritchard, 1982). Pritchard (1982)
used aerial surveys to assess an area of Mexico that resulted in discovery of a large,
previously unknown leatherback rookery. With this new information, Pritchard (1982)
was able to create a crude estimate of the world's population of mature, female
leatherback turtles. Sarti et al. (1996) followed up the work of Pritchard (1982) and other
researchers, and found this large Mexican conglomeration of nesting leatherbacks to be in
serious decline from previously documented numbers. Crouse (1984) conducted aerial
surveys in North Carolina in conjunction with ground truthing, allowing for a conversion
factor to create a seasonal estimate for the state of North Carolina. Unfortunately,
differential crawl retention on individual beaches precluded any season-long estimate.
Aerial surveys have been flown over the TTI area, but documentation of sea turtle
nesting was not the primary objective of the flights. Irvine et al. (1981) flew fixed-wing
manatee and dolphin surveys over the TTI, documenting incidental observations of sea
turtle crawls in the area. LeBuff and Hagan (1978) flew fixed-wing aerial surveys over
the beaches of Cape Romano and Marco Island, and reported missing approximately 50%
of the nests documented by surveyors on the ground. Fixed-wing aerial surveys by Davis
and Whiting (1977) over the Cape Sable region extended into the southern portion of the
Ten Thousand Islands archipelago, encompassing Pavilion and Rabbit Keys, included in
the current study. Kushlan (1986) repeated Davis and Whiting's surveys 8 years later,
and found results similar to the previous work.
The objective of this study was to develop a seasonal nesting estimate for the TTI
using an aerial survey coupled with ground truth counts. The estimate developed
depended upon two basic assumptions: nesting activity on the 8 ground-truth islands is
similar to activity on the other study islands and the nest-to-non-nesting emergence ratio
on other islands in the TTI is similar to that experienced on the 8 ground-truth islands.
Materials and Methods
The Ten Thousand Islands (TTI) extend roughly 40 kilometers from Marco Island
to Pavilion Key along Florida's southwest coast (Edwards, 1991). Located
approximately 25 kilometers southeast of Naples, these largely undeveloped islands are
utilized by the public year-round, primarily for fishing (Cheng, 1996). The climate is
subtropical with 110 to 160 centimeters of average annual precipitation, two-thirds of
which occurs between May and October (Cheng, 1996). Average annual temperature is
23 degrees Celsius, with occasional freezing temperatures in the winter and regular mid-
thirties in the summer (Cheng, 1996).
These islands are dominated by mangrove trees, with some invasion by Australian
pine (Casuarina equisetifolia). Beaches are characterized by sand, false coral, or oyster
shells, either individually or in some combination on each island. Characteristically,
beaches of these islands face the Gulf of Mexico from many different aspects and are
much shorter and narrower than the long, ribbon-like beaches on the east coast of Florida
or in Naples, Florida.
From aerial photographs and fly-overs prior to initiation of the study, 29 keys in
the TTI were identified that could potentially support sea turtle nesting. From this
information, a flight plan was developed that allowed observation of all 29 keys for
sufficient numbers of days to create an overall seasonal nesting estimate for the TTI in
1996. To assess the accuracy of the aerial counts for the 29 islands, 8 of these islands
were chosen as ground truth islands for this study. Sea turtle nesting surveys were
conducted from May 13, 1996 until August 15, 1996 for the 8 ground-truth islands.
Ground surveys were conducted twice a week for each island, providing a seasonal
nesting total. This strategy was developed by the U.S. Fish and Wildlife Service due to
the remoteness of the TTI and lack of personnel.
Aerial surveys were conducted in a U.S. Coast Guard Sikorsky Jayhawk
helicopter at an altitude between 46.9 to 62.5 meters and a speed of 30 to 40 knots over
the surf zone. This placement of the helicopter afforded the best view of the narrow TTI
beaches. Flights were made on two successive mornings after new moons in June (June
19 and 20) and July (July 9 and 10) and on two successive mornings after full moons in
June (June 26 and 27) and July (July 25 and 26). The flight on June 19 was a training
flight, and the flight on June 27 was canceled due to mechanical problems with the
The full moon/new moon flight plan provided the greatest proportion of "fresh"
crawls (i.e. crawls made the night prior to aerial survey), since the tides following these
lunar cycles erase most old crawls from the beaches. In addition, all old crawls on the
ground-truth beaches were erased by raking prior to aerial survey. Crawls were ground-
truthed by assistants conducting daily surveys to provide the basis for a visibility index
and an accurate nest/false crawl ratio. Aerial surveys were begun between 0800 and
0900 h, usually concluding within two hours, beginning at Coon Key at the northwest end
of the study area, and ending at Pavilion Key at the extreme southeastern end. The
survey was conducted from the port side of the helicopter. This methodology was
adapted from research conducted by Hopkins-Murphy and Murphy (1983) in South
The following formula was used to calculate a nesting estimate for the entire
If the total ground and aerial counts for ground-truth islands are not equal, then a
correction factor has to be calculated:
G = ground count of crawls for ground-truth islands
Ag = aerial count of crawls for ground-truth islands
C = correction factor
The correction factor can then be applied to the observed number of aerial counts
to create a corrected estimate (a) of the total number of aerial counts at all sites:
Ta = total aerial count of crawls at all sites
= total corrected # of aerial counts at all sites
If a correction factor is necessary then the total aerial estimate (a) would then be
substituted for the total aerial count (Ta) in subsequent calculations. The percentage of
total nesting activity on the ground-truth islands observed during the aerial survey is
calculated in order to obtain a seasonal nesting estimate:
Tg = total seasonal ground count of crawls for ground-truth islands
P = % of total seasonal crawl activity for ground-truth islands
Once the percentage of the total seasonal nesting activity (P) has been calculated,
a total seasonal nesting estimate can be calculated:
S= total seasonal crawl estimate for TTI
After the total seasonal nesting estimate has been calculated, the number of total
nests can be calculated based on the nest-to-total-crawl ratio for the 8 ground-truth
Ng = # of nests observed on ground-truth islands during aerial survey
f = total # of nests for all sites
A total of 26 crawls was observed in both ground and aerial surveys during the
study period on the ground-truth islands (Table 3-1). Although the total counts were
identical, the overall aerial survey failed to detect one crawl on one occasion and
recorded one when it did not exist on another occasion. In addition, the ground and aerial
counts for individual islands exhibited a much higher degree of variability when
compared to the total count for the entire survey period (Table 3-2).
Table 3-1. Ground and aerial counts of sea turtle crawls on 8 islands in TTI for June-July
Date Ground count Aerial count
6/20/96 5 5
6/26/96 8 8
7/9/96 5 4
7/10/96 4 4
7/25/96 3 4
7/26/96 1 1
TOTAL 26 26
Table 3-2. Aerial and ground counts on ground-truth islands for
each survey by island
Date Key Ground Air
7/9/96 Brush 7
Date Key Ground
The total number of crawls observed based on ground surveys conducted twice a
week throughout the 1996 nesting season (Figure 3-1), does not accurately represent the
levels of activity on the 8 ground truth islands on any given night. Figure 3-2 depicts the
estimated numbers of nests that were laid on the 8 ground truth islands, based on the
twice-a-week ground surveys. These nightly estimates were calculated by totaling the
number of nests laid during lapses in ground surveys and dividing that total by the
number of nights in the interval between ground surveys. The average obtained was then
applied to each of the nights that fell between the interval between ground surveys. This
graph (Figure 3-2) represents a trend in the amount of nightly nesting activity that
occurred on these islands during the 1996 season.
An estimate was derived utilizing the formulas described previously in this paper.
There were 26 crawls (G) counted on the 8 ground-truth islands during the aerial survey.
For these 8 ground-truth islands, there were a total of 422 crawls (Tg) in 1996. Thus, the
aerial observations accounted for 6.2% (P) of the seasonal crawl activity for these 8
islands. Since there were 118 aerial observations (Ta) at all sites, a total seasonal crawl
estimate (8) of 1903 crawls was derived. By calculating the % of nests for the 8 ground-
truth islands (Ng/Tg= 47.6%) for the entire nesting season, the total number of nests at all
sites (u = 906) was calculated. It is estimated that 906 nests and 997 non-nesting
emergences occurred on the 29 study islands during the 1996 nesting season.
Aerial surveys have been used in many sea turtle studies (Shoop et al., 1985;
Crouse, 1984; Stancyk et al., 1979; Fritts et al., 1982; Meylan and Meylan, 1985;
Pritchard, 1982) as a technique for covering large areas of coastline, reaching remote
areas and reducing costs of conducting season-long daily or nightly surveys. Aerial
surveys have been employed on several occasions in the area immediately adjacent to the
L I I *h' U) =
1 01 0 0 10 0N
peiuesqo SIMR-- -o .Ioq fnN
In 0 I n
peAjesqo sIM Jo j joeqwunN
* a g
' a >
' K s
' p: a
' R s
' R R
^ R Fi
^ R S
' K a
' ^ R
'. r:: H
' K a
' -= a
' R "
' ft s
' 1= ?
' K S
' R a
TTI (Davis and Whiting, 1977; LeBuff, Jr. and Hagan, 1978; Irvine, 1981; Kushlan,
1986), reporting varying degrees of success at documenting sea turtle activity.
The most efficient means of gathering data on marine turtle nesting activity over a
large area appears to be aerial survey (Hopkins-Murphy and Murphy, 1983). But, aerial
survey is only useful if it is effective, that is, producing accurate and reproducible
estimates (Hopkins-Murphy and Murphy, 1983). There are a number of factors that
influence the effectiveness of aerial survey nesting estimates. LeBuff and Hagan (1978)
reported that strong westerly or northwesterly winds in conjunction with high tides often
caused beach erosion and shifting sands, and that severe weather left marine vegetation
upon beaches that concealed crawls. Hopkins-Murphy and Murphy (1983) discuss the
importance of aerial sampling parameters associated with aerial surveys such as: speed of
the aircraft, height of the aircraft, position of the aircraft, angle of view (beach), skill of
the pilot and observer experience on estimator accuracy. Other factors that could
potentially bias aerial estimates, include: amount of turtle activity, the type of beach, the
amount of sunlight, human activity on the beach, location of crawls on the beach and in
the case of the TTI, the ordinal aspect of the islands. LeBuff and Hagan (1978) noted
that crawls in close proximity to each other or crawls partially hidden caused problems
when attempting to assess nesting activity from the air. Not all factors can be controlled
or accounted for in creating nesting estimates.
Crouse (1984) identified one of the inherent difficulties with creating estimates
from aerial survey, is that a limited number of aerial surveys are used to extrapolate to a
total seasonal nesting estimate. Crouse (1984) is critical of the use of a conversion factor
as a means of creating nesting estimates via aerial survey. She noted that variability of
crawl retention and visibility from beach to beach adversely affects conversion indices.
These factors may vary from beach to beach depending upon characteristics of the
nesting substrate and prevailing tides (Crouse, 1984). She also stated that weather
patterns could affect crawl retention, both geographically and temporally, and that
changes in beach profile due to erosion or accretion may affect crawl retention over time
The TTI are used minimally in the summer months by humans, therefore, that
impact on crawl retention is minimal. In addition weather patterns and the type of
nesting substrate can cause differential crawl retention on TTI beaches temporally; these
factors can be controlled by only documenting "fresh" crawls. Recognizing these
limitations of aerial survey, efforts to reduce potential sources of bias and make
corrections prior to the beginning of the flight schedule, were taken. For example, a test
flight was conducted where fine-tuning of the aerial survey technique to the TTI took
place (i.e. adjusting speed, altitude and angle of observation of the helicopter). Another
problem associated with aerial survey is detecting all crawl activity after nights of high
nesting activity or storm activity. One of the methods of dealing with this problem is the
creation of a conversion factor. Although the overall accuracy of a seasonal estimate will
suffer if aerial and ground counts are not in fairly close agreement, the conversion factor
still allows for an estimate to be made. The accuracy of this sort of estimate would be
debatable, although it would provide managers with a good idea of the trends in sea turtle
nesting activity for their study area. The usefulness of such data cannot be discounted.
Since the survey technique suggested here has numerous limitations, it is
necessary to describe methods that could improve upon the protocol used for the 1996
season. One of the key limitations of this particular study is the fact that we conducted 6
flights. After assessing the data, it was determined that in order to accurately create a
nesting estimate that is worthy of application for threatened and endangered species,
several years of flights would need to be flown, with the flights occurring no less than a
daily basis for at least the first year. After collecting this data, it would be possible to
determine the minimum number of flights needed to create estimates that can limit the
variability in the data. By adding many more flights to the protocol, a more accurate
picture of sea turtle nesting trends could potentially be developed for the TTI.
Although the technique applied in the TTI for the 1996 nesting season is not the
answer to developing an accurate nesting estimate for the TTI, it did provide answers
about the difficulties associated with aerial survey in the TTI. Therefore, the technique
served as the first step in potential development of a method for accurately assessing sea
turtle nesting activity in the TTI. Furthermore, the nesting estimate suggested in this
paper should not be viewed as a hard estimate of the total activity in the TTI. In fact, due
to the variability of crawls observed, the estimate should be cautiously viewed as an
indicator of the amount of sea turtle nesting that is occurring in the TTI. This preliminary
research should be expanded upon by implementing the changes to the methodology
suggested in this paper.
Davis and Whiting (1977), in work that is now over 20 years old, found an
estimated 1553 loggerhead nests in 1972 and 919 loggerhead nests in 1973 for nesting
beaches stretching from Cape Sable northwest to Morman Key, just southeast of the
furthest extent (Pavilion Key) of my study area. Kushlan (1986) conducting aerial
surveys in the same area and employing the same methodology as Davis and Whiting
(1977) found 1362 nests in 1980. Taking into account the estimate I made for the TTI,
this portion of coastline, long thought to be fairly insignificant in terms of overall nesting
effort by sea turtles, could potentially host between two to three thousand sea turtle nests
in a given year. These nesting numbers identify the significance of this coastline as a
high use area for sea turtles, particularly the loggerhead sea turtle. Furthermore, I was
unable to differentiate crawls by species of sea turtle from the aerial survey, therefore, it
is not possible to say how many of the nests in the TTI are from green turtles (Chelonia
mydas), which are known to nest in the islands.
This technique may be used as an indicator of nesting activity in the TTI
although, there are several additional changes in the methodology used in this study that
could potentially increase the accuracy of aerial surveys. The use of two observers is a
strategy used in the work conducted by Murphy and Hopkins-Murphy (1983) in South
Carolina. They found no significant difference in the number of crawl identifications
between two separate observers. By using two observers, like Murphy and Hopkins-
Murphy (1983), a basic mark-recapture study could be created, and the results could be
subjected to statistical analyses to lend strength to any aerial estimates. Since aerial
observations are conducted by helicopter in the TTI, there is greater flexibility with
regard to how flights could be conducted. It may be possible to safely reduce the speed at
which the helicopter flies, based on the discretion of the pilot. It may be beneficial to
hover over beaches at the discretion of the observerss. Hovering might allow for more
accurate documentation of crawl activity. The observers) may request additional fly-bys
of beaches in order to get more accurate counts of crawl activity.
The methodology suggested here is imperfect. Although the estimates for 1996
appear to be sound, these results cannot predict the accuracy of aerial surveys for future
years in the TTI. Therefore, when implementing this strategy, the conclusions drawn
from the results presented should be considered cautiously.
Aerial survey, in conjunction with seasonal ground surveys, appears to be a useful
tool for evaluation of sea turtle activity in the Ten Thousand Islands. In order to develop
low-cost methods for assessing trends in sea turtle nesting in the TTI, natural resource
managers may consider annual aerial surveys with partial ground-truth counts punctuated
with aerial survey corrected by season-long ground surveys. Detailed information on the
accuracy of aerial survey, as well as the trends in sea turtle nesting in the TTI may allow
for the development of a lower-cost method for evaluating nesting activity. However,
before any protocol can be established, at least several more years of aerial surveys need
to be conducted, before nesting trends can be more clearly understood.
MOVEMENT PATTERNS OF THE TEN THOUSAND ISLANDS RACCOON
(PROCYONLOTOR MARINUS), AND THE EFFECT OF ITS REMOVAL ON SEA
TURTLE HATCH SUCCESS
Introduction and Background
Raccoons (Procyon lotor) have been the subject of considerable study pertaining
to various aspects of behavior, feeding ecology, and population dynamics. In addition,
several studies have dealt directly with the Ten Thousand Islands raccoon (Procyon lotor
marinus). This subspecies was identified by Goldman (1950), and later confirmed by
Hall and Kelson (1959). This spurred further study, including descriptions of
morphometric characteristics (Bigler et al, 1977), and population characteristics of P. 1.
marinus (Bigler et al, 1981).
Raccoons are effective predators of sea turtle nests in Florida (Ratnaswamy, 1995;
McMurtray, 1986; Lewis et al., 1994; Kushlan, 1986; Klukas, 1967; Davis & Whiting,
1977), Georgia (Anderson, 1981), and South Carolina (Stancyk et al., 1980; Andre &
West, 1979; Hopkins & Murphy, 1980; Hopkins et al., 1977). Raccoon depredation rates
of sea turtle nests ranged from 49-87% annually at Cape Sable, FL (Davis & Whiting,
1977) and from 59-85% annually from 1991-1994 for Ten Thousand Islands National
Wildlife Refuge (TTINWR), FL (Garmestani, 1995). Raccoons have been identified as
the primary predator of sea turtle nests with humans, fire ants, ghost crabs and bobcats as
possible secondary predators in TTINWR (L. Richardson, pers. comm.).
Due to the imperiled status of sea turtles, various means of decreasing raccoon
depredation of sea turtle nests have been explored. Numerous studies have been
conducted to evaluate the effectiveness of conditioned taste aversion (CTA) and its
impacts on consumption of eggs by raccoons. Linhart et al. (1991) formulated baits that
were readily taken by raccoons, tested the baits in the field, and found that 31-53% of the
baits were taken by non-target species (Linhart et al., 1994). Conover (1990) reported
that emetine dihydrochloride was determined to have some effectiveness in reducing egg
depredation. Nicolaus (1987) and Semel & Nicolaus (1992) found CTA using estrogen-
injected eggs can be an effective strategy to reduce raccoon depredation of eggs.
However, Nicolaus et al (1989), found mixed results when treating eggs with estrogen,
and researchers in South Carolina found CTA using lithium chloride to be completely
ineffective at reducing depredation of sea turtle nests (Hopkins & Murphy, 1982). One
study conducted in South Dakota concluded that raccoons could be baited into ingesting
chemosterilants, thereby inhibiting reproduction (Nelson, 1970). CTA techniques were
rejected by TTINWR because of their mixed results, combined with the likelihood of
ingestion by non-target species.
Other methods of sea turtle nest protection have been attempted. McMurtray
(1986) concluded that the most effective means of turtle nest protection was flat screens
left on nests until hatchlings emerged. Ratnaswamy (1995) supported the claim that
screens are the most effective form of nest protection from predators, and claimed that
CTA and predator removal were ineffective. Several researchers argue the effectiveness
of nest transplantation as an effective means of decreasing turtle egg mortality (Andre &
West, 1979; Stancyk et al., 1980). Hopkins and Murphy (1982) concluded that the most
effective form of nest protection includes nest transplantation and predator removal,
thereby decreasing chances of inundation and depredation. This conclusion was
supported by the findings of researchers at St. Vincent's National Wildlife Refuge (Lewis
et al., 1994). However, the U.S. Fish and Wildlife Service has recently determined that
nest transplantation should only be considered when a nest is threatened by potential
inundation (MacPherson, personal comm.). Klukas (1967) reported a 25% reduction in
nest destruction using predator elimination as a management tool on Cape Sable in
Everglades National Park (ENP).
Raccoons depredated nests on the eight continuously surveyed islands in
TTINWR at varying rates (Garmestani, 1995). Although the literature indicated that
other forms of nesting management may be more effective, particularly screening in
conjunction with predator removal (Hopkins & Murphy, 1982; Lewis et al., 1994), lack
of personnel and the remoteness of TTINWR precluded any attempt at labor-intensive or
season-long nesting management. Due to these constraints, the effectiveness of short-
term raccoon control was assessed as a possible management strategy for the Ten
Materials and Methods
Study areas were 4 low-relief mangrove islands located within TTINWR and the
Cape Romano-Ten Thousand Islands Aquatic Preserve in southwest Florida. This rich
estuary, which supports breeding and feeding grounds for a multitude of species, consists
of a labyrinth of tidal creeks, bays, and passes. The 4 study islands were dominated by
red (Rhizophora mangle), black (Avicennia germinans), and white (Laguncularia
racemosa) mangrove trees, interspersed with sandy and/or shelly beaches. These beaches
are of varying lengths (50m-815m) and widths (1.4m-18.3m), and support beach
vegetation that varies with the substrate and elevation of the particular island and/or
beach. Study islands were of varying sizes, in close proximity (< 2 km) and
interconnected by the observed motility of raccoons: Panther (54.8 ha), White Horse
(39.3 ha), Hog (26.7 ha), and Gullivan (19.6 ha). These 4 islands (Panther, Hog, White
Horse, and Gullivan Keys) were known to support sea turtle nesting and to suffer varying
rates of raccoon depredation of sea turtle nests (Garmestani, 1995). White Horse was the
site of 75 documented sea turtle nests from 1991 through 1994, and had an average
depredation rate of 46% for those four years; Gullivan was the site of 124 documented
sea turtle nests from 1991-1994, with a 91% depredation rate for that time period; Hog
Key had a total of 21 documented sea turtle nests for the years 1991-1994, with an
average depredation rate of 72%.
Live traps (91.44cm X 27.94cm X 27.94cm [2-door]; 81.28cm X 30.48cm X
25.4cm [1-door]; Havahart, Inc., Lititz, Pennsylvania, USA; 121.92cm X 30.48cm X
30.48cm [1-door]; custom traps, South Carolina Wildlife and Marine Resources
Department, Georgetown, South Carolina, USA) were placed in a randomly spaced array
in the vicinity of nesting beaches, utilizing vegetation as camouflage. Raccoons were
attracted into live traps using a variety of baits, including sardines, a sardine/cat food
mixture, mullet, marshmallows, and chicken eggs coated with fish oil. After a raccoon
was captured, it was compressed within the live trap using either a homemade plunger or
stakes. The animal was then injected intramuscularly with 8-10 mg/kg of ketamine (100
mg/ml; Fort Dodge Laboratories, Inc., Fort Dodge, Iowa, USA). Once the animal was
anesthetized, one of two things then occurred: 1) if the animal was part of the mark-
recapture study it received a Monel (National Band and Tag Co., Newport, KY, USA) ear
tag in each ear and a Passive Integrated Transponder (PIT) (Avid Co., Norco, CA, USA)
which was placed in the scruff of the neck, and a 4 cc sample of blood (1995) was taken
from the heart and placed on ice; 2) if the animal was part of the removal study, blood
was drawn (1995), and the animal euthanized with an intercardiac injection of euthasol
(Delmarvia Laboratories, Inc., Midlothian, VA, USA). Next, numerous aspects of the
anatomy of the raccoon were measured and evaluated, including: sex, testes lengths,
baculum length, evidence of lactation, weight, age (Grau et al., 1970), canine lengths,
total length (head-body and tail), right hind-foot, right ear, length of vibrissae, neck, head
and chest circumference, pelage, and physical condition. Euthanized raccoons were then
buried in three-foot deep holes on the beach. Tagged raccoons were released by placing
them in the shade, covering their eyes with horseshoe crab shells to alleviate the potential
damage from sunlight. Raccoons were captured and removed on Panther Key in January,
February, May, June, and July for 1995, and in January for 1996; sampled by mark-
recapture on White Horse Key in March of 1995 and 1996, and on Hog Key in April of
1995 and 1996; Gullivan Key was sampled by mark-recapture in March and April of
1995, and removal in March and April of 1996.
1995 Field Season
On January 11th, 1995, raccoon trapping commenced on Panther Key in the
TTINWR. A 100 mile round trip drive, and a 12 mile round trip boat ride were required
each day in order to arrive at the field site in the TTINWR. Initially, 12 Havahart one-
door live traps were used to capture raccoons on the west beach of Panther Key.
However, after compiling 8 nights using only the one-door traps, 12 Havahart two-door
live traps were deployed on the east beach of Panther Key for an additional 15 nights, for
a total of 23 nights on Panther Key. Only one door was opened on the two-door traps
each night. Beginning May 1, 12 one-door and 10 custom live traps on loan from South
Carolina Wildlife and Marine Resources Department (SCWMRD) were re-set on Panther
Key. After 5 nights of trapping, the 10 custom traps were removed, and the remaining 12
live traps were run until July 14, 1995. Raccoons were enticed into traps with a variety
of baits including sardines, a sardine/cat food mixture, marshmallows and mullet.
Beginning March 3, 1995, raccoon trapping commenced on White Horse Key. A
total of 33 traps was initially deployed on White Horse, including 12 Havahart one-door
traps, 11 Havahart two-door traps, and 12 custom traps on loan from Phillip Wilkinson of
SCWMRD. Six of the two-door traps were closed after two nights of trapping, because
of time constraints. All raccoons captured were marked and released. Mullet was used
exclusively to bait traps on White Horse Key.
Beginning March 28, 1995, raccoon trapping commenced on Gullivan Key.
Twenty-two traps were set on Gullivan, including 12 Havahart one-door traps, and 10 of
the custom traps on loan from SCWMRD. Mullet was used exclusively as bait, and all
raccoons were tagged and released.
Beginning April 17, 1995, 7 one-door Havahart traps were set on Hog Key. This
number was upgraded the following day, by adding 5 more Havahart one-door traps, and
10 of the custom traps on loan from SC Wildlife. A total of 8 nights were run on Hog
Key from April 17 through April 28. Mullet and marshmallows were used as bait, and all
captured raccoons were tagged and released.
1996 Field Season
Raccoon trapping for the 1996 field season was based upon the rate of first-time
captures for the 1995 field season (Table 4-1).
Table 4-1. Initial raccoon capture rates for TTINWR
(1995) initial trapping regime
Island new captures # of nights %new captures
Panther 13 8 100
White Horse 18 5 100
Gullivan 8 6 100
Hog 10 5 100
In addition, the study methodology was further altered from the 1995 field season
by adding Gullivan Key as a removal island, along with Panther Key. Gullivan Key was
selected due to its distance from Panther Key and high depredation rate. Also, during
1996, mark-recapture efforts were conducted on White Horse and Hog Keys, in order to
accumulate data on changes in raccoon population characteristics of the two islands.
Raccoon trapping commenced on Panther Key beginning on January 18, 1996,
running traps for 8 nights. Twenty-one traps were deployed on the island, including 11
Havahart one-door and 10 Havahart two-door live traps. Each of these traps was baited
with chicken eggs scented with fish oil. This bait was devised as a method of targeting
egg-depredating raccoons. All raccoons captured were euthanized and buried on site.
On March 5, 1996, trapping began on White Horse Key, running 5 nights.
Twenty traps were originally deployed on White Horse, including 20 Havahart live traps
(e.g. ten one-door, ten two-door). But, after two nights, one of the two-door traps was
stolen, so the remaining three nights were run with 19 traps. All traps were baited with
the chicken egg/fish oil bait, and all raccoons were marked and released.
On March 30, 1996, trapping commenced on Gullivan Key, running 6 nights.
Nineteen Havahart live traps were deployed on Gullivan, including 10 one-door and 9
two-door traps. All traps were baited with egg/oil, and all raccoons were eliminated and
buried on site.
On April 18, 1996, trapping was initiated on Hog Key, running 5 nights. Eighteen
Havahart live traps were deployed on Hog Key, including 10 one-door and 8 two-door
traps. All traps were baited with egg/oil, and all raccoons were marked and released.
1995 Field Season
Panther Key (PK), an island continuously surveyed for the past 5 years at
TTINWR, was chosen as a test island for raccoon removal because of its high incidence
of sea turtle nests and raccoon depredation. In addition to the limited success with live
trapping, baiting the beach with piles of mullet was completely unsuccessful for spot-
lighting and shooting of raccoons. There were a total of 15 captures on Panther Key,
including one escape. So, a total of 14 raccoons was removed from Panther Key in 750
trap nights for the 1995 season. Traps were reset on Panther Key in May to determine
whether raccoons will move from adjacent islands to exploit an open niche and a plentiful
seasonal food resource (i.e. sea turtle nests). However, only 1 unmarked raccoon was
captured after the start of sea turtle nesting. Raccoon tracks were observed on numerous
occasions on Panther Key during turtle nesting season, but no depredation occurred in
1995 (Table 4-2).
Table 4-2. Sea turtle nesting and depredation on Panther Key
for the 1991-1995 nesting seasons
Year total nests # depredated % depredated
1991 63 63 100
1992 42 40 95
1993 36 29 81
1994 38 29 76
1995* 41 0 0
Even with the elimination of raccoon depredation as a factor in sea turtle nest
mortality for Panther Key in 1995, overall hatch success did not significantly improve
due to the high rate of inundation of nests (Table 4-3). This high rate of inundation was
due to an extreme aberration in the occurrence of tropical storms and hurricanes for the
1995 storm season.
Table 4-3. Sea turtle hatch success on Panther Key for the 1995 nesting season
In addition to the animals trapped on Panther Key, raccoons were marked and
released on Gullivan (n=8), White Horse (n=18), and Hog (n=10) Keys. All 3 of these
islands are in close proximity to Panther Key, and all have turtle nesting beaches that
have been documented for the past 5 years.
Ten nights of trapping were run on White Horse Key with 18 initial captures, and
a total of 30 captures for White Horse Key. Raccoons were tagged and released,
accounting for the 12 recaptures (Table 4-4). A total of 12 nights of trapping was run on
Gullivan Key, with 8 initial captures and 1 recapture (Table 4-4). Ten initial captures
were recorded on Hog Key in 8 nights of trapping (Table 4-4).
Table 4-4. Results of raccoon trapping activity on the four study islands for the
1995 trapping season
# of trap Initial Total # of captures % of
Island nights captures recaptures per night successful
Panther 750 14 0 0.33 2.0
White 286 18 12 3.00 10.5
Gullivan 264 8 1 0.75 3.4
Hog 161 10 1 1.38 6.8
TOTAL 1461 50 14 0.67 3.4
Of these 50 captures, 36 were males, ranging in size from 1.4 kg to 4.8 kg, with mean
weight and standard error of 3.55 +/- 0.136 kg. Fourteen were females ranging in size
from 1.4 kg to 3.6 kg, with mean weight and standard error of 2.83 +/- 0.151 kg. Age
class distribution for males (n=36) captured in 1995 was fairly even (Figure 4-1), but age
class distribution was slightly skewed towards young (Class 1) and middle-aged (Class 3)
females (n=14) (Figure 4-1). Although a population estimate could not be created,
raccoon densities were calculated for the period of trapping on a particular island (Table
sOo0e NO #
SU0033Rej Jo #
Table 4-5. Raccoon densities during trapping periods
Island # of raccoons captured Density during capture
period (per ha.)
Panther 14 0.26
White Horse 18 0.46
Gullivan 8 0.37
Hog 10 0.41
TOTAL 50 0.36
1996 Field Season
Raccoon removal was conducted again in 1996 on Panther Key. There were a
total of 2 captures in 168 trap nights. No depredation was observed in 1996, and I
assume that this is attributable to the removal of raccoons on Panther Key (Table 4-6).
Table 4-6. Number of depredated nests for Panther Key
Year # of nests # depredated % depredated
1996 61 0 0
The hatch success for sea turtles on Panther Key drastically improved for the
1996 field season when compared to hatch rates for the 1995 field season (Table 4-7). I
attribute this drastic increase to the effectiveness of raccoon removal and the lower
incidence of devastating storms (i.e. hurricanes and tropical storms) in 1996.
Table 4-7. Hatch success for Panther Key (1996)
# of nests # hatched # of eggs # of emerged # unhatched
61 58 (95%) 5809 4875 (84%) 934 (16%)
In addition to the animals trapped on Panther Key, raccoons were removed on
Gullivan Key (n=l), and mark-recapture was conducted on White Horse (n=15) and Hog
There were 15 initial captures and 1 recapture for a total of 16 captures on White
Horse for 1996. Of these 15 captures, 6 were captured on White Horse in 1995 and 1 was
tagged on Gullivan in 1995. This male raccoon is the only record of movement between
islands based on my mark-recapture data. Therefore, 8 "new" raccoons were trapped and
tagged on White Horse Key in 1996. All raccoons were tagged and released.
There was 1 capture on Gullivan for 1996. This individual was not one of the 8
raccoons tagged on Gullivan in 1995. The individual was eliminated and buried on-site.
The raccoon removal method was completely ineffective for Gullivan in 1996. Raccoons
depredated nests at a similar rate in 1995, when no removal was attempted (Table 4-8).
Table 4-8. Number of depredated nests for Gullivan Key
Year # of nests # depredated % depredated
1995 36 34 94
1996* 33 32 97
Five nights of trapping were run on Hog Key, from April 18-22, 1996. There
were a total of 6 captures on Hog, including 1 recapture. Of these 5 initial captures, 2
were captured in 1995, and 3 were "new" raccoons. All raccoons were marked and
Of these 23 captures, 13 were males, ranging in size from 2.1 kg to 4.7 kg, with
mean weight and standard error of 3.45 +/- 0.182 kg, and 10 were females ranging in size
from 1.6 kg to 3.6 kg, with mean weight and standard error of 2.64 +/- 0.208 kg. Age
class distribution for females (n= 10) captured in 1996 was fairly even (Figure 4-2), but
age class distribution was slightly skewed towards young (Class 1) and middle-aged
(Class 3) males (n=13) (Figure 4-2). Trapping data are summarized in Table 4-9.
Table 4-9. Raccoon activity in TTINWR (1996)
# of trap Initial Total # of captures % of
Island nights captures recaptures per night successful
Panther 168 2 0 0.25 1.2
White 97 15 1 3.00 16.5
Gullivan 114 1 0 0.17 0.8
Hog 108 5 1 1.2 5.6
TOTAL 487 23 2 1.04 5.1
Although a population estimate could not be created for 1996, raccoon densities during
trapping periods are summed in Table 4-10.
Table 4-10. Raccoon densities during trapping periods
Island # of raccoons captured Density during capture
period (per ha.)
Panther 2 0.04
White Horse 15 0.38
Gullivan 1 0.05
Hog 5 0.04
TOTAL 23 0.16
The data I collected provide a closer look at the Ten Thousand Islands raccoon
(Procyon lotor marinus), which has been studied by several other researchers (Nelson,
1930; Bigler et al., 1977; Bigler et al., 1977). Although I was able to capture quite a few
raccoons (1995= 50; 1996= 23), I was unable to get enough recaptures to run any sort of
population model, such as the examples suggested in Hallett et al. (1991). Therefore, I
was unable to create a population estimate for the 4 islands I surveyed.
However, I was able to collect information pertaining to the natural history of P. 1.
marinus. For instance, in 1995, I caught more males (n= 36) than females (n= 14). It is
possible that this male to female ratio is due to a sex-biased response to live traps as has
been suggested by Gehrt and Fritzell (1996), who found that males were captured more
frequently during all of their trapping periods. Yet, in 1996, although I have a small
sample size, I caught raccoons at nearly a 1:1 ratio (males n= 13, females n= 10). This
ratio (1:1) is what is to be expected in nature, although the exploration of this sex-bias
theory merits a closer look.
Depredation of sea turtle nests on Panther Key was reduced to zero for the two-
year study period. I assume this drastic decrease in depredation on Panther Key was due
to raccoon control. This finding is inconclusive, but unconnected research conducted on
St. Catherine's Island in Georgia found that there was no shift in raccoon home ranges or
centers of activity with regard to sea turtle nesting (Anderson, 1981). Wilkinson (pers.
comm.) observed in South Carolina that raccoons not familiar with sea turtle nests may
forage on beaches without taking eggs obviously available for exploitation.
Finally, when considering the poor trap success for the 1996 field season, the
intangible factor of trap-smart raccoons must be taken into account. Short of conducting a
radio-telemetry study, it would be difficult to quantify what percentage of the TTI
raccoons are trap-smart, although there is evidence to prove their persistence in the
population. For example, low-capture rates on Hog Key and to a lesser degree White
Horse Key were experienced in 1996 after those islands were subjected to trapping in
1995. But the greatest evidence is provided by the abysmal trap success on Gullivan Key
coupled with extremely high nest depredation for the island (see Table 4-8 and Figure 4-
3). From the limited mark-recapture data, only 1 raccoon moved from 1 island to
another. This evidence and the fact that low numbers of raccoons re-populated Panther
Key, suggest that raccoon removal could be an effective sea turtle management policy for
the Ten Thousand Islands (Figure 4-4).
The Ten Thousand Islands raccoon has proven to be the single greatest threat to
sea turtle hatch success in the mangrove forests of southwest Florida. Therefore, any sea
turtle management strategy should include live trapping on islands that have not
previously been exposed to trapping, and explore the possibility of utilizing leg-hold
25 --Nests Laid
20 -.- Depredated
1991 1992 1993 1994 1995 1996
Figure 4-3. Raccoon depredation rates on Gullivan Key, 1991-1996.
Raccoon control program in effect this year
1991 1992 1993 1994 1995 1996
1991 1992 1993 1994 1995 1996
Figure 4-4. Raccoon depredation rates on Panther Key, 1991-1996.
Raccoon control program in effect these years
traps and shooting as methods of removal for islands that have been exposed to live
trapping. In addition, removal efforts should be concentrated on islands with high
numbers (>10) of sea turtle nests and/or with the highest probability of hatch success, due
to slope, sediment type, etc. These islands include: B, Gullivan, White Horse and Hog
Keys. Finally, hatch success will be dramatically affected by weather patterns for the
area. High hatch success cannot be expected if numerous storms have direct and indirect
effects on the islands as in 1995, a year of heavy storm activity, and 1996, a year of
minimal storm activity (see Table 4-3 and 4-7).
It would be beneficial to know about the movement patterns and re-population
trends for the TTI raccoons. Conducting radio-telemetry on P.I. marinus may provide
answers that allow for a long-term raccoon management strategy to be employed. By
conducting this study, information could be gathered that may indicate how often raccoon
control measures would need to be implemented. It would also be necessary to evaluate
the population genetics of each islands' raccoon population. Understanding the
relatedness of raccoons on each of the islands, coupled with radio-telemetry, would
provide a more complete understanding of the raccoon populations and their movement
patterns in the TTI, and allow for a more comprehensive raccoon management strategy.
It appears that loggerhead sea turtles select nesting beaches according to certain
criteria of beaches in the TTI. The conclusions reached about loggerheads in the TTI
may not be true of C. caretta in other parts of the world. Other factors may play a larger
part in the decision-making process in sea turtles. For instance, beach lighting, the
presence of humans, sea walls and re-nourished beaches could potentially have a greater
effect on nesting beach and nest-site selection, particularly in inhabited areas, than the
factors I analyzed for the TTI. However, the most important element of my study is that
the nesting habitat in the TTI is relatively unspoiled. This means that the nesting
preferences exhibited by loggerheads can be assumed to be based on cues inherent to the
1) Aerial surveys can be effective in assessing sea turtle activity in the TTI,
reducing expenditures of time, personnel and expenses. Obviously aerial surveys
need to be conducted in a helicopter due to the configuration of the islands, but
with continued cooperation of the U.S. Coast Guard, an effective, low-cost survey
plan could be implemented for the region. The use of aerial survey also provided
the first comprehensive evaluation of sea turtle nesting activity for the TTI. This
season-long estimate may fluctuate with sea turtle nesting cycles, and changes to
the survey technique, but this indicator of sea turtle nesting activity in the TTI
should be considered a minimum nesting estimate for the region when you take
into account the five islands that were not included in this study, including one
island (Cape Romano) that is known to support from 50 to 100 nests per season
(Kraus, pers. comm.).
2) The removal of raccoons is a potential management strategy for improvement
of sea turtle hatch success. Even if depredation is reduced significantly, hatch
success will not necessarily improve in years of heavy storm activity in the area.
Therefore, raccoon removal is simply a method for decreasing depredation of sea
turtle nests and indirectly increasing hatch success. In addition, the raccoon
research provided some basic data on the status of Procyon lotor marinus in the
3) Other research addressing the effects of exotic invasive species (i.e. Australian
pine and the red imported fire ant) into the TTI needs to be undertaken to
understand their impacts on sea turtles, as well as other native flora and fauna in
the archipelago. In addition, personal observations of Kemp's Ridley
(Lepidochelys kempii) sea turtles and juvenile green turtles in the area warrants
investigation of this area as a potential feeding ground for these two species.
Addison, D.A. 1994. Caretta caretta (loggerhead sea turtle) nesting. Herpetological
Review 25: 63.
Allard, M.W., M.M. Miyamoto, K.A. Bjorndal, A.B. Bolten and B.W. Bowen. 1994.
Support for natal homing in green turtles from mitochondrial DNA sequences.
Copeia 1994: 34-41.
Anderson, S. 1981. The raccoon (Procyon lotor) on St. Catherines Island, Georgia: 7.
Nesting sea turtles and foraging raccoons. American Museum Novitates 2713: 1-
Andre, J. B., and L. West. 1981. Nesting and management of the Atlantic loggerhead,
Caretta caretta caretta (Linnaeus) (Testudines: Cheloniidae) on Cape Island,
South Carolina, in 1979. Brimleyana No. 6: 73-82.
Bayliss, P. 1986. Factors affecting aerial surveys of marine fauna, and their relationship
to a census of dugongs in the coastal waters of the Northern Territory. Australian
Wildlife Resources 13: 27-37.
Bigler, W.J., A.S. Johnson, and G.L. Hoff. 1977. Morphometric characteristics of the
Ten Thousand Islands raccoon. Florida Scientist 40: 261-262.
Bigler, W.J., G.L. Hoff, and A.S. Johnson. 1981. Population characteristics of Procyon
lotor marinus in estuarine mangrove swamps of southern Florida. Florida
Scientist 44: 151-157.
Bjorndal, K.A., A. Carr, A.B. Meylan and J.A. Mortimer. 1985. Reproductive biology
of the hawksbill (Eretmochelys imbricata) at Tortuguero, Costa Rica, with notes
on the ecology of the species in the Caribbean. Biological Conservation 34: 353-
Bjorndal, K.A. and A.B. Bolten. 1992. Spatial distribution of green turtle (Chelonia
mydas) nests at Tortuguero, Costa Rica. Copeia 1992: 45-53.
Bowen, B.W., A.B. Meylan, J.P. Ross, C.J. Limpus, G.H. Balazs and J.C. Avise. 1992.
Global population structure and natural history of the green turtle (Chelonia
mydas) in terms of matriarchal phylogeny. Evolution 46: 865-881.
Bowen, B.W., J.C. Avise, J.I. Richardson, A.B. Meylan, D. Margaritoulis and S.
Hopkins-Murphy. 1993. Population structure of the loggerhead turtle (Caretta
caretta) in the northwest Atlantic Ocean and Mediterranean Sea. Conservation
Biology 7: 834-844.
Bowen, B.W. 1995. Molecular genetic studies of marine turtles. In Biology and
Conservation of Sea Turtles, K.A. Bjorndal, Ed.-Rev. Ed., pp. 585-587.
Broderick, D., C. Moritz, J.D. Miller, M. Guinea, R.J. Prince and C.J. Limpus. 1994.
Genetic studies of the hawksbill turtle: evidence for multiple stocks and mixed
feeding grounds in Australian waters. Pacific Conservation Biology 1: 121-131.
Bustard, H.R. 1972. Australian Sea Turtles. London: Collins and Sons.
Bustard, H.R. 1979. Population dynamics of sea turtles. In Turtles: Perspectives and
research: Marion Harless & Henry Morlock, editors; 523-540, 1979.
Camhi, M.D. 1993. The role of nest site selection in loggerhead sea turtle (Caretta
caretta) nest success and sex ratio control. Ph.D. Dissertation, Rutgers
University, New Brunswick, NJ. 255pp.
Carr, A. and M.H. Carr. 1972. Site fixity in the Caribbean green turtle. Ecology 53:
Carr, A., M.H. Carr and A.B. Meylan. 1978. The ecology and migration of sea turtles. 7.
The west Caribbean turtle colony. Bulletin of the American Museum of Natural
History 162: 4-46.
Chaloupka, M.Y. and J.A. Musick. 1997. Age, growth, and population dynamics. In
The Biology of Sea Turtles, P.L. Lutz and J.A. Musick, Eds., pp. 233-276.
Cheng, E. 1996. Marine turtle report. Ten Thousand Islands National Wildlife Refuge,
Naples, FL. Unpublished. Mimeo. 27pp.
Christens, E. and J.R. Bider. 1987. Nesting activity and hatching success of the painted
turtle (Chrysemys picta marginata) in southwestern Quebec. Herpetologica 43:
Christiansen, J.L. and B.J. Gallaway. 1984. Raccoon removal, nesting success, and
hatchling emergence in Iowa turtles special reference to Kinosternonflavescens
(Kinosternidae). The Southwestern Naturalist 29: 343-348.
Conover, M.R. 1990. Reducing mammalian predation on eggs by using a conditioned
taste aversion to deceive predators. Journal of Wildlife Management 54: 360-
Crouse, D.T. 1984. Loggerhead sea turtle nesting in North Carolina: applications of an
aerial survey. Biological Conservation 29: 143-155.
Crouse, D.T., L.B. Crowder and H. Caswell. 1987. A stage-based population model for
loggerhead sea turtles and implications for conservation. Ecology 68: 1412-
Daud, S.K. and M.W.R.N. De Silva. 1987. Nesting site selection of leatherback and
green turtles on Terengganu Beach. Proceedings of the 10th Annual Seminar of
the Malaysian Society of Marine Sciences, pp.78-83.
Davis, G.E. and M.C. Whiting. 1977. Loggerhead sea turtle nesting in Everglades
National Park, USA. Herpetologica 33: 18-28.
Dean, Jr., W.E. 1974. Determination of carbonate and organic matter in calcareous
sediments and sedimentary rocks by loss of ignition: comparison with other
methods. Journal of Sedimentary Petrology 44: 242-248.
Edwards, K. 1991. Marine turtle nesting report. Ten Thousand Islands National
Wildlife Refuge, Naples, FL. Unpublished. Mimeo. 23pp.
Ehrenfeld, D.W. and A.L. Koch. 1967. Visual accommodation in the green turtle. Science
Fangman, M.S. and K.A. Rittmaster. 1994. Effects of human beach usage on the
temporal distribution of loggerhead nesting activities. Proceedings of the 13th
Annual Symposium on Sea Turtle Biology and Conservation, pp. 222-224.
Fowler, L.E. 1979. Hatching success and nest predation in the green turtle, Chelonia
mydas, at Tortuguero, Costa Rica. Ecology 60: 946-955
Frazer, N.B. 1983. Demography and life history evolution of the Atlantic loggerhead sea
turtle, Caretta caretta (Georgia). Ph.D. Dissertation, University of Georgia,
Frazer, N.B. 1983. Effect of tidal cycles on loggerhead sea turtles (Caretta caretta)
emerging from the sea. Copeia 1983: 516-519.
Frazer, N.B. 1984. A model for assessing mean age-specific fecundity in sea turtle
populations. Herpetologica 40: 281-291.
Frederick, P.C. and M.W. Collopy. 1989. The role of predation in determining
reproductive success of colonially nesting wading birds in the Florida Everglades.
The Condor 91: 860-867.
Fritts, T.H., M.L. Stinson and R. Marquez M. 1982. Status of sea turtle nesting in
southern Baja California, Mexico. Bulletin of the Southern California Academy
of Sciences 81: 51-60.
Fritts, T.H., W. Hoffman and M.A. McGehee. 1983. The distribution and abundance of
marine turtles in the Gulf of Mexico and nearby Atlantic waters. Journal of
Herpetology 17: 327-344.
Garmestani, A.S. 1995. Island biogeography: Raccoon movement patterns with respect
to a seasonal food resource. U.S. Fish and Wildlife Service, Naples, FL. Mimeo.
Gehrt, S.D. and E.K. Fritzell. 1996. Sex-biased response of raccoons (Procyon lotor) to
live traps. American Midland Naturalist 135: 23-32.
Goldman, E.A. 1950. Raccoons of north and middle America. North American Fauna
60. U.S. Fish and Wildlife Service. Washington, D.C.
Grassman, M.A., D.W. Owens, J.P. McVey and R. Marquez M. 1984. Olfactory-based
orientation in artificially imprinted sea turtles. Science 224: 83-84.
Grau, G.A., G.C. Sanderson and J.P. Rogers. 1970. Age determination of raccoons.
Journal of Wildlife Management 34: 364-372.
Hall, E.R. and K.R. Kelson. 1959. The mammals of North America. Vol. 2: 884-891.
Ronald Press Co. New York.
Hallett, J.G., M.A. O'Connell, G.D. Sanders and J. Seidensticker. 1991. Comparison of
population estimators for medium-sized mammals. Journal of Wildlife
Management 55: 81-93.
Harris, M.J. and C.G. Maley. 1990. Loggerhead sea turtle nesting on Ossabaw, St.
Catherines, Sapelo, and Little St. Simons Island, Georgia in 1989. U.S. Fish and
Wildlife Service Study Completion Report Project E-4, Study I: 28pp.
Hartman, L.H. 1993. Ecology of coastal raccoons (Procyon lotor) on the Queen
Charlotte Islands, British Columbia, and evaluation of their potential impact on
native seabirds. M.S. Thesis, University of Victoria, Vancouver, 155pp.
Hays, G.C. and J.R. Speakman. 1993. Nest placement by loggerhead turtles, Caretta
caretta. Animal Behaviour 45: 47-53.
Hays, G.C., A. MacKay, C.R. Adams, J.A. Mortimer, J.R. Speakman and M. Boerema.
1995. Nest site selection by sea turtles. Journal of Marine Biological
Assessments (U.K.). 75: 667-674.
Henwood, T.A. 1987. Sea turtles of the southeastern United States, with emphasis on
the life history and population dynamics of the loggerhead turtle, (Caretta
caretta). Ph.D. Dissertation, Auburn University, Auburn, AL. 170pp.
Hopkins, S.R., T.M. Murphy, Jr., K.B. Stansell, and P.M. Wilkinson. 1978. Biotic and
abiotic factors affecting nest mortality in the Atlantic loggerhead turtle.
Proceedings of the Annual Conference of the Southeastern Association of Fish
and Wildlife Agencies 32: 213-223.
Hopkins, S.R., and T.M. Murphy, Jr. 1982. Management of loggerhead turtle nesting
beaches in South Carolina. South Carolina Wildlife and Marine Resources
Department Report E-1, Study No. VI-A-2, 19pp.
Hopkins, S.R., and T.M. Murphy, Jr. 1982. Testing of lithium chloride aversion to
mitigate raccoon depredation of loggerhead turtle nests. Proceedings of the
Annual Conference of the Southeastern Association of Fish and Wildlife
Agencies 36: 484-491.
Hopkins-Murphy, S.R. and T.M. Murphy, Jr. 1983. Distribution of loggerhead turtle
nesting activity in South Carolina by aerial beach survey. South Carolina
Wildlife and Marine Resources Department Report E-1, Study No. VI-A-2, 60pp.
Hopkins-Murphy, S.R., and T.M. Murphy, Jr. 1985. Factors affecting nest survival of
Caretta caretta in South Carolina: implications for management. Unpublished,
Hughes, G.R. 1974. The sea turtles of southeast Africa. 1. Status, morphology and
distributions. Oceanographic Research Institute Investigational Report, Durban
Irvine, A.B., J.E. Caffin and H.I. Kochman. 1981. Aerial surveys for manatees and
dolphins in western peninsular Florida (with notes on sightings of sea turtle and
crocodiles). U.S. Fish and Wildlife Service, Washington, D.C. FWS/OBS-80/50,
Johannes, R.E. and D.W. Rimmer. 1984. Some distinguishing characteristics of the
green turtle (Chelonia mydas) on North West Cape Peninsula, Western Australia.
Marine Biology 83: 149-154.
Kilham, L. 1986. Renestings of American crows in Florida and predation by raccoons.
Florida Field Naturalist 14: 21- 23.
Klukas, R.W. 1967. Factors affecting nesting success of loggerhead turtles at Cape
Sable, Everglades National Park. File Number N1415 National Park Service, PO
Box 279, Homestead, Florida, 33033. 58pp.
Koch, A.L., A. Carr and D.W. Ehrenfeld. 1969. The problem of open-sea navigation:
the migration of the green turtle to Ascension Island. Journal of Theoretical
Biology 22: 163-179.
Kushlan, J.A. 1986. Atlantic loggerhead nesting status in southwest Florida.
Herpetological Review 17: 51- 52.
LeBuff, Jr., C.R. 1969. The marine turtles of Sanibel and Captiva islands. Sanibel-
Captiva Conservation Foundation, Inc., Report No.l: 1-13.
LeBuff, Jr., C.R. and R.W. Beatty. 1971. Some aspects of nesting of the loggerhead
turtle, Caretta caretta caretta (Linne), on the gulf coast of Florida. Herpetolgica
LeBuff, Jr., C.R. and P.D. Hagan. 1978. The role of aerial surveys in estimating nesting
populations of the loggerhead turtle. Florida Marine Research Publications 33:
LeBuff, Jr., C.R. 1990. The loggerhead turtle in the eastern Gulf of Mexico. Caretta
Research, Inc., Sanibel, FL, 236pp.
Lewis, T.E., G.O. Bailey, H.L. Edmiston. 1994. Effects of predator control on sea turtle
nest success on the barrier islands of Apalachicola Bay. Proceedings of the 14th
Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical
Memorandum NMFS-SFESC-351: 242-243.
Linhart, S.B., F.S. Blom, G.J. Dasch, J.D. Roberts, R.M. Engeman, J.J. Esposito, J.H.
Shaddock, and G.M. Baer. 1991. Formulation and evaluation of baits for oral
rabies vaccination of raccoons (Procyon lotor). Journal of Wildlife Diseases 27:
Linhart, S.B., F.S. Blom, R.M. Engeman, H.L. Hill, T. Hon, D.I. Hall, and J.H.
Shaddock. 1994. A field evaluation of baits for delivering oral rabies vaccines to
raccoons (Procyon lotor). Journal of Wildlife Diseases 30: 185-194.
Lohmann, K.J. 1992. How sea turtles navigate. Scientific American January: 100-106.
Lund, P.F. 1986. Nest production and nesting-site tenacity of the loggerhead turtle,
Caretta caretta, on Jupiter Island, Florida. Unpublished M.S. Thesis, University
of Florida, Gainesville. 32pp.
Maley, C.G. and M.J. Harris. 1991. Loggerhead sea turtle nesting on Ossabaw, St.
Catherines, Sapelo, and Little St. Simons Island, Georgia in 1990. U.S. Fish and
Wildlife Service Study Completion Report Project E-4, Study I: 36pp.
Maley, C.G. and M.J. Harris. 1992. Loggerhead sea turtle nesting on Ossabaw, St.
Catherines, Sapelo, and Little St. Simons Island, Georgia in 1991. U.S. Fish and
Wildlife Service Study Completion Report Project E-4, Study I: 37pp.
Maley, C.G. and M.S. Murphy. 1993. Loggerhead sea turtle nesting on Ossabaw, St.
Catherines, Sapelo, and Little St. Simons Island, Georgia in 1992. U.S. Fish and
Wildlife Service Study Completion Report Project E-4, Study I: 23pp.
Maley, C.G. and M.S. Murphy. 1994. Loggerhead sea turtle nesting on Ossabaw, St.
Catherines, Sapelo, and Little St. Simons Island, Georgia in 1993. U.S. Fish and
Wildlife Service Study Completion Report Project E-4, Study I: 21pp.
Maley, C.G. 1995. Loggerhead sea turtle nesting on Ossabaw, St. Catherines, Sapelo,
and Little St. Simons Island, Georgia in 1994. U.S. Fish and Wildlife Service
Study Completion Report Project E-4, Study I: 16pp.
Marsh, H. and D.F. Sinclair. 1989. An experimental evaluation of dugong and sea turtle
aerial survey techniques. Australian Wildlife Research 16: 639-650.
McMurtray, J.D. 1986. Reduction of raccoon predation on sea turtle nests at Canaveral
National Seashore, Florida. M.S. Thesis, University of Georgia, Athens. 48pp.
Melucci, C.L., J.I. Richardson, R. Bell and L.A. Corliss. 1991. Nest site preference and
site fixity of hawksbills on Long Island, Antigua. Proceedings of the 11th Annual
Workshop on Sea Turtle Biology and Conservation, pp. 171-174.
Meylan, A. and P. Meylan. 1985. Nesting of Dermochelys coriacea in Caribbean
Panama. Journal of Herpetology 19: 293-297.
Meylan, A.B., B.W. Bowen and J.C. Avise. 1990. A genetic test of the natal homing
versus social facilitation models for green turtle migration. Science 248: 724-
Miller, J.D. 1995. Nesting biology of sea turtles. In Biology and Conservation of Sea
Turtles, K.A. Bjorndal, Ed.-Rev. Ed., pp.573-575.
Mortimer, J.A. and K.M. Portier. 1989. Reproductive homing and interesting behavior
of the green turtle (Chelonia mydas) at Ascension Island, South Atlantic Ocean.
Copeia 1989: 962-977.
Mortimer, J.A. 1990. The influence of beach sand characteristics on the nesting
behavior and clutch survival of green turtles (Chelonia mydas). Copeia 1990:
Mortimer, J.A. 1995. Factors influencing beach selection by nesting sea turtles. In
Biology and Conservation of Sea Turtles, K.A. Bjorndal, Ed.-Rev. Ed., pp. 45-52.
Mourao, G.M., P. Bayliss, M.E. Coutinho, C.L. Abercrombie and A. Arruda. 1994. Test
of an aerial survey for caiman and other wildlife in the Pantanal, Brazil. Wildlife
Society Bulletin 22: 50-56.
Mrosovsky, N. 1983. Ecology and nest-site selection of leatherback turtles,
Dermochelys coriacea. Biological Conservation 26: 47-56.
Murphy, T.M. and S.R. Hopkins-Murphy. 1990. Homing of translocated gravid
loggerhead turtles. Proceedings of the 10th Annual Workshop on Sea Turtle
Biology and Conservation, pp. 123-124.
Nelson, E.W. 1930. Four new raccoons from the keys of southern Florida. Smithsonian
Miscellaneous Collections 82: 1-17.
Nelson, R.L. 1970. Evaluation of chemosterilants as inhibitors of reproduction in
raccoons. South Dakota Department of Game, Fish and Parks progress report W-
75-R- 12: 10pp.
Nicolaus, L.K. 1987. Conditioned aversions in a guild of egg predators: implications
for aposematism and prey defense mimicry. Department of Biological Sciences,
Northern Illinois University, DeKalb 60115.
Nicolaus, L.K., J. Herrera, J.C. Nicolaus, and C.R. Gustavson. 1989. Ethinyl estradiol
and generalized aversions to eggs among free-ranging predators. Applied Animal
Behaviour Science 24: 313-324.
Pollock, L.W. and W.D. Hummon. 1971. Cyclic changes in interstitial water content,
atmospheric exposure, and temperature in a marine beach. Limnology and
Oceanography 16: 522-535.
Portier, K.M. and D.L. Anderson. 1995. Using tree regression to identify nutritional and
environmental factors affecting sugarcane production. Applied Statistics in
Agriculture 7: 182-189.
Pritchard, P. 1971. The leatherback or leathery turtle, Dermochelys coriacea. IUCN
Monograph, Marine Turtle Series 1: 1-39.
Pritchard, P.C.H. 1982. Nesting of the leatherback turtle, Dermochelys coriacea, in
Pacific Mexico, with a new estimate of the world population status. Copeia 1982:
Ratnaswamy, M.J. 1995. Raccoon depredation of sea turtle nests at Canaveral National
Seashore, Florida: implications for species management and conservation. Ph.D.
Dissertation, University of Georgia, Athens. 79pp.
Richardson, J.I. 1982. A population model for adult female loggerhead sea turtles
(Caretta caretta) nesting in Georgia. Ph.D. Dissertation, University of Georgia,
Richardson, J.I., T.H. Richardson and M.W. Dix. 1978. Population estimates for nesting
female Loggerhead sea turtles (Caretta caretta) in the St. Andrew Sound area of
southeastern Georgia. Florida Marine Research Publications 33: 34-38.
Richardson, T.H., J.I. Richardson and M.W. Dix. 1978. Remigration patterns of
loggerhead sea turtles (Caretta caretta) nesting on Little Cumberland and
Cumberland Islands, Georgia. Florida Marine Research Publications 33: 39-44.
Sarti, M.L., S.A. Eckert, T.N. Garcia and A.R. Barragan. 1996. Decline of the world's
largest nestingassemblage of leatherback turtles. Marine Turtle Newsletter 74:
SAS Institute, Inc. 1989. SAS/STAT User's guide. SAS Institute, Cary, NC. 846pp.
Schulte, E.E. 1988. Recommended soil organic matter tests. In Recommended chemical
soil test procedures for the north-central region, W.C. Dhanke, Ed. North Dakota
Agricultural Experiment Station, North Dakota State University Bulletin #499:
Schulz, J.P. 1975. Sea turtles nesting in Surinam. Stichting Natuurbehoud Suriname
(STINASU) 3: 1-143.
Seigel, R.A. 1979. Predation by raccoons on diamondback terrapins, Malaclemys
terrapin tequesta. Journal of Herpetology 14: 87-89.
Semel, B. and Nicolaus, L.K. 1992. Estrogen-based aversion to eggs among free-
ranging raccoons. Ecological Applications 2: 439-449.
Shabica, S., D. Stoneburner, D. Gilmore, J. Hedden, J. Hinesley and J. Swigert. 1978.
Raccoon predation on sea turtle nests. Bulletin of the Ecological Society of
America 59: 102.
Shoop, C.R., C.A. Ruckdeschel and N.B. Thompson. 1985. Sea turtles in the southeast
United States: nesting activity as derived from aerial and ground surveys, 1982.
Herpetologica 41: 252-259.
Stancyk, S.E. and J.P. Ross. 1978. An analysis of sand from green turtle nesting beaches
on Ascension Island. Copeia 1978: 93- 99.
Stancyk, S.E., O.R. Talbert and A.B. Miller. 1979. Estimation of loggerhead turtle
nesting activity in South Carolina by aerial survey. American Zoologist 19: 954.
Stancyk, S.E., O.R. Talbert and J.M. Dean. 1980. Nesting activity of the loggerhead
turtle (Caretta caretta) in South Carolina: II. Protection of nests from raccoon
predation by transplantation. Biological Conservation 18: 289-298.
Statistical Sciences. 1993. S-PLUS User's manual. StatSci, Seattle.
Stoneburner, D.L. and J.I. Richardson. 1981. Observations on the role of temperature in
loggerhead turtle nest site selection. Copeia 1981: 238-241.
Thompson, N.B. 1980. Population dynamics of the Atlantic green sea turtle, Chelonia
mydas, (Linnaeus) 1758. Ph.D. Dissertation, University of Rhode Island,
Urban, D. 1970. Raccoon populations, movement patterns, and predation on a managed
waterfowl marsh. Journal of Wildlife Management 34: 35-39.
Venables, W.N. and B.D. Ripley. 1994. Modern Applied Statistics with S-PLUS.
Springer-Verlag, New York. 462pp.
Westin, L. and A. Nissling. 1991. Effects of salinity on spermatozoa motility,
percentage of fertilized eggs and egg development of Baltic cod (Gadus morhua),
and implications for cod stock fluctuations in the Baltic. Marine Biology 108: 5-
Whitmore C.P. and P.H. Dutton. 1985. Infertility, embryonic mortality and nest-site
selection in leatherback (Dermochelys coriacea) and green sea turtles (Chelonia
mydas) in Surinam. Biological Conservation 34: 251-272.
Witherington, B.E. 1995. Hatchling orientation. In Biology and Conservation of Sea
Turtles, K.A. Bjorndal, Ed.-Rev. Ed., pp. 577-578.