Origin of juvenile loggerhead turtles (Caretta caretta) in a 
tropical developmental habitat in Caribbean Panam?
INTRODUCTION
The migratory lifestyle and temporally protracted life
history of marine turtles present difficulties for the effec-
tive conservation of these species (Carr, Carr & Meylan,
1978; Bowen, 1995). Hatchlings leave the nesting beach
immediately after emerging from the nest, swim directly
offshore, and enter a pelagic life-history stage for 10 or
more years prior to recruiting as juveniles to near-shore
developmental habitats (Carr, 1986, 1987; Bolten &
Balazs, 1995). Turtles either may remain in the same
developmental habitat through maturation (Limpus,
Couper & Read, 1994a,b), or they may migrate among
different developmental habitats and then migrate to a
separate adult feeding ground where maturation takes
place (Carr et al., 1978; Ehrhart, 1989; A. Meylan & P.
Meylan, unpublished data). Adult turtles make periodic
reproductive migrations from feeding grounds to nest-
ing beaches that may be separated by tens, hundreds or
thousands of kilometers (Meylan, 1982; Limpus et al.,
1992; Bowen et al., 1995). Despite subdivision of pop-
ulations on the nesting beaches, extensive migration of
post-hatchling, juvenile and adult turtles provides oppor-
tunity for these reproductively isolated stocks to mix in
foraging habitats, thus making it difficult to assess how
exploitation of turtles on feeding grounds will affect
reproductive populations (Carr et al., 1978; Bowen &
Karl, 1997).
All sea turtle nesting populations examined to date
show a pattern of strong population subdivision in mito-
chondrial DNA (mtDNA). This pattern is consistent with
the natal homing hypothesis (Meylan, Bowen & Avise,
1990) and also suggests that genetically distinct nesting
aggregations are demographically independent and must
be treated as separate management units (Bowen & Karl,
1997). Differences in mtDNA haplotype frequencies
among nesting populations provide genetic markers that
are powerful tools for the identification of the natal ori-
gin of turtles captured away from nesting beaches
(Bowen & Karl, 1997). Maximum likelihood based tech-
niques of mixed stock analysis (Pella & Milner, 1987)
are now widely applied to estimate nesting beach
Tag N. Engstrom1,3, Peter A. Meylan1 and Anne B. Meylan2
1 Natural Sciences Collegium, Eckerd College, 4200 54th Ave S., St. Petersburg, FL 33711, USA 
2 Florida Fish and Wildlife Conservation Commission, Florida Marine Research Institute, 100 8th Avenue SE, St. Petersburg, FL 33701,
USA
3 Present address: Section of Evolution and Ecology and Center for Population Biology, University of California, One Shields Aven.,
Davis, CA 95616, USA
(Received 16 April 2001; accepted 26 November 2001)
Abstract
Maximum likelihood mixed stock analysis was used to identify the natal origin of immature logger-
head turtles (Caretta caretta) in a tropical developmental habitat in Caribbean Panam?. Approximately
65?70% of the loggerhead turtles in Chiriqu? Lagoon originate from South Florida nesting beaches,
and the other 30?35% originate from Mexico. Haplotype frequencies of the Chiriqu? Lagoon logger-
head population are significantly different from those observed in the pelagic environment in the east-
ern Atlantic, and estimated nesting beach contributions to Chiriqu? Lagoon are significantly different
from values expected if recruitment were based solely on the size of nesting populations. These obser-
vations suggest that dispersal of loggerheads into benthic developmental habitats from the pelagic
environment is not random. The occurrence of US and Mexican loggerheads in tropical developmental
habitats has not been previously recognized. Exploitation and other mortality factors operating in the
Caribbean area must be taken into account in demographic models and management plans for these
two populations. This exploitation could be particularly important for the small, demographically vul-
nerable Mexican population and for other small populations for which no genetic data are currently
available.
All correspondence to: Tag N. Engstrom. 
Section of Evolution and Ecology, University of California, 
One Shields Ave, Davis, California 95616. Tel.: 530?752?1112;
Fax: 530?752?1449; E-mail: tnengstrom@ucdavis.edu.
Animal Conservation (2002) 5, 125?133  ? 2002 The Zoological Society of London 
DOI:10.1017/S1367943002002184 Printed in the United Kingdom
contributions to marine turtle feeding grounds (Bowen
et al., 1996; Bass, Lageuex & Bowen, 1998; Bolten et
al., 1998; Laurent et al., 1998; Bass & Witzell, 2000;
Rankin-Baransky et al., in press). This information has
provided valuable insights into previously intractable
questions regarding the ecological geography of marine
turtles and has been particularly useful in tracking migra-
tion, dispersal and recruitment patterns of juvenile tur-
tles. Information about the origin of turtles on foraging
grounds has direct applications to conservation because
incidental and direct take of turtles on foraging grounds
has the potential to affect reproductive populations on a
wide geographic scale (Bowen & Karl, 1997). 
Since 1987, the Bocas Turtle Project has studied the
ecology and migrations of the immature green (Chelonia
mydas) loggerhead (Caretta caretta) and hawksbill
turtles (Eretmochelys imbricata) that use the eastern end
of the Chiriqu? Lagoon as developmental habitat
(A. Meylan & P. Meylan, unpublished data). Chiriqu?
Lagoon is an approximately 2000 km2 bay located on
the Caribbean coast of Panam? in Bocas del Toro
Province, at 09o 00? N, 081o 50? W (Fig. 1). No major
nesting beaches for loggerhead turtles currently exist in
the tropics of the western Atlantic, so the occurrence of
immature loggerheads in this part of the tropical Atlantic
seems anomalous, and the origin of the turtles in Chiriqu?
Lagoon has been completely unknown. In this compo-
nent of the larger long-term study, we use sequence data
from the mtDNA control region to estimate the contri-
bution of Atlantic Ocean stocks to the Chiriqu? Lagoon
foraging ground. We use our results to test predictions
regarding patterns of recruitment of juvenile marine tur-
tles to developmental habitats, and we evaluate the
potential that ongoing exploitation in Chiriqu? Lagoon,
although small-scale, may adversely affect small, vul-
nerable nesting populations that are geographically far
removed.
MATERIALS AND METHODS
Turtles for this study were captured in the eastern end
of Chiriqu? Lagoon using a netting technique that is com-
mon in the turtle fisheries throughout the Caribbean
region. Large-mesh (25?40 cm) tangle nets 10?80 m
long are held in place at each end by an anchor and a
balsa wood buoy. In lieu of a lead line, the net is held
vertical in the water by tying a series of small rocks to
the bottom mesh, thus allowing an entangled turtle to
breathe by bringing the net to the surface. Entangled tur-
tles were removed from nets at least twice daily and
brought back to a base camp on shore for measurement
and sample collection. 
The straight carapace length (notch-to-notch, standard
and maximum) and width, curved carapace length,
plastron length and length of tail to the cloaca and to the
tip were measured using standard techniques as
described by Bolten (1999). Laparoscopic examination
of the gonads was used to determine the sex and maturity
of 16 loggerhead turtles captured during 1989?1991 and
1997 following the criteria of Limpus & Reed (1985).
In subsequent years maturity was assessed by size. Blood
samples for genetic analysis were collected from the 45
immature loggerhead turtles captured in Chiriqu? Lagoon
during May?June field seasons of 1993, 1994 and 1997.
Blood was drawn from the cervical sinus as described
by Owens & Ruiz (1980) and preserved in a lysis buffer
consisting of 10 mM EDTA, 100mM Tris-HCl and 1.0%
SDS at pH 8.0 at a blood:buffer ratio of 1:10. Blood
samples were stored at ambient temperature in the field
until transport to the laboratory where they were stored
at 4oC. Genomic DNA was isolated from blood samples
using standard phenol:chloroform extraction methods
(Palumbi, 1996) and stored at ?20o C. 
A ~390 base pair (bp) segment of mtDNA control
region was amplified by the polymerase chain reaction
(PCR) (Saiki et al., 1986) using the primer pair TCR5
(5?-TTGTACATCTACTTATTTACCAC-3?) TCR6 (5?-
GTACGTACAAGTAAAACTACCGTATGCC-3?)
(Norman, Moritz & Limpus, 1994). Standard precau-
tions including the use of negative controls were taken
to avoid contamination. PCR products were purified
using either strepdaviadin-coated beads (Megabeads,
Promega Corp., Madison, WI), or 30,000 ng DNA fil-
ters (Ultrafree-MC Millipore Corp., Bedford, MA) and
sequenced at the DNA Sequencing Core Lab at the
University of Florida or at the Division of Biological
Sciences DNA Sequencing Facility at the University of
California, Davis. Sequences were aligned by eye with
previously published loggerhead control region
sequences (Bolten et al., 1998; Encalada et al., 1998)
and compared with haplotypes in the online haplotype
registry maintained by the Archie Carr Center for 
Sea Turtle Research (ACCSTR http://accstr.ufl.edu).
Throughout this paper we use standardized haplotype
nomenclature established by ACCTR but also include
the letter code of the previous convention in parenthe-
ses where appropriate. Sequences that showed ambigu-
ity or did not match a previously published sequence
were sequenced in both the forward and reverse direc-
tion. We tested for significant differences in haplotype
frequencies among samples from Chiriqu? Lagoon,
pelagic habitats (Bolten et al., 1998) and nesting beaches
(Encalada et al., 1998; Laurent et al., 1998) by using
Monte Carlo simulation as implemented in the program
Monte Carlo RxC (W. Engles, University of Wisconsin). 
Encalada et al. (1998) identified six genetically dis-
tinct stocks among loggerhead turtle nesting colonies in
the Atlantic and Mediterranean. Laurent et al. (1998)
identified an additional Mediterranean stock in Turkey
and examined 60 more turtles from throughout the south-
eastern Mediterranean. Differences in haplotype fre-
quencies between Encalada?s and Laurent?s samples
from Greece and the southeastern Mediterranean were
not significant (P > 0.25) so all are combined as a sin-
gle Greece?southeastern Mediterranean stock in our
analyses. In this paper we sometimes refer to this
Greece?SE Med population and the genetically distinct
Turkey nesting population collectively as the
Mediterranean populations. We use the names estab-
lished by the Turtle Expert Working Group (TEWG,
126 T. N. ENGSTROM ET AL.
127
O
rigin of juvenile loggerheads in Panam?
Northern (NEFL-NC)
South Florida
Mexico
Florida panhandle
Major loggerhead rookeries
Smaller unsampled rookeries      
Chiriqu? Lagoon
Bastimentos Island
National Marine Park
Brazil
Chiriqu?
Lagoon
Net sites
Fig. 1. Map of the western Atlantic showing the location of loggerhead turtle nesting beaches in relation to Chiriqu? Lagoon, Panam?. Locations of major
nesting populations are indicated with the figure of a turtle, which is roughly proportional to the size of the nesting population (see Table 3). Locations of
smaller populations for which no mtDNA information is available are indicated with question marks. Sites of turtle capture are indicated by dots on the
inset map of Chiriqu? Lagoon. The current boundary of the Bastimentos Island National Marine Park is outlined with a broken line
2000) for nesting populations in the Atlantic and Gulf
of Mexico.
We assumed equal contributions as a starting point for
iterations of two different maximum likelihood mixed
stock analyses: (1) including only Atlantic nesting
beaches as possible source populations; (2) including
both Atlantic and Mediterranean nesting populations as
source populations. In each of these two analyses, rela-
tive contributions of nesting populations in the
Atlantic?Mediterranean basin to the Chiriqu? Lagoon
foraging ground were estimated by using three different
mixed stock analysis programs: CONSQRT, UCON
(Masuda, Nelson & Pella, 1991) and SHADRACQ (Xu,
Kobak & Smouse, 1994). Population estimates for each
nesting population in the Atlantic?Mediterranean basin
(Laurent et al., 1998; TEWG, 2000) were used to cal-
culate nesting beach contributions that would be
expected under a model of random recruitment. These
expected contributions were compared with our esti-
mated contributions by using adjusted X2 goodness-of-
fit test (Sokal & Rohlf, 1995). 
RESULTS
Since 1987, a total of 81 loggerheads have been cap-
tured in Chiriqu? Lagoon. The only mature Caretta
captured at this site was a 98-cm adult male, which had
recently lost its entire right front flipper. The remaining
80 turtles ranged in size from 45.5 to 76.5 cm straight
carapace length notch-to-notch (SCLN), with a mean
length of 59.8 cm (N = 80, SD = 6.91), and were judged
to be immature on the basis of laparoscopy (N = 16),
small size, and the lack of development of secondary sex
characteristics. The 45 turtles examined in this genetic
study ranged from 48.1 to 73.3 cm SCLN with a mean
length of 60.1 cm (N = 45, SD = 7.07).
Aligned nucleotide sequences for 380 bp of the
mtDNA control region from these turtles yielded a total
of 25 polymorphic sites consisting of 22 transitions, no
transversions and three insertion/deletions (indels).
These polymorphic sites defined eight haplotypes among
the 45 Chiriqu? Lagoon turtles. Six of the haplotypes
observed in Chiriqu? Lagoon match haplotypes reported
from nesting beaches (Encalada et al., 1998; Laurent et
al., 1998). Two turtles from Chiriqu? Lagoon showed
haplotypes that have not yet been observed on any nest-
ing beach. Haplotype CC-A12 (L) was previously found
in a single pelagic-stage loggerhead (Bolten et al., 1998);
CC-A19 was reported from one stranded loggerhead
from Georgia (A. L. Bass and B. W. Bowen, unpub-
lished data). Haplotype CC-A12 differs from the
endemic Mexican haplotype, CC-A10 (J), by a single
transition at position 63. Haplotype CC-A19, differs
from the common CC-A2 (B) haplotype by a single tran-
sition at position 259. Haplotype frequencies in Chiriqu?
Lagoon did not differ across years (P = 0.56) and thus
were pooled for all analyses (Table 1).
In pair-wise comparisons, haplotype frequencies in
Chiriqu? Lagoon differ significantly from those of each
individual nesting population (P < 0.05 in all cases)
except for South Florida (P = 0.19). However, the pres-
ence of two endemic Mexican haplotypes, CC-A9 (I)
and CC-A10 (J), in our sample indicates that turtles from
the Mexican population are also present in Chiriqu?
Lagoon. Haplotype frequencies in Chiriqu? Lagoon also
differ significantly from those observed in the pelagic
habitat in the eastern North Atlantic (P = 0.043) (Bolten
et al., 1998). 
Because the mixed stock analysis programmes used
here required that all individuals in the mixed popula-
tion be assigned to types occurring in the included source
populations, our analysis was limited to the 43 turtles
with haplotypes already identified from nesting beaches
(Encalada et al., 1998; Laurent et al., 1998). In analy-
ses including only Atlantic nesting beaches as possible
contributing sources (left columns of Table 2), contri-
128 T. N. ENGSTROM ET AL.
Table 1. Distribution of mtDNA haplotypes of loggerhead turtles (Caretta caretta) in Chiriqu? Lagoon, Panam? (developmental habitat), the
eastern Atlantic (pelagic habitat), and seven Atlantic and Mediterranean nesting populations. Haplotypes occurring in Chiriqu? Lagoon are
indicated with an asterisk
.
Haplotype Chiriqu? Eastern Florida South Northern Mexico Brazil Greece? Turkey
letter code Lagoon Atlantic panhandle Florida NEFL-NC SE Med
CC-A1* 10 60 34 22 104
CC-A2* 26 50 4 24 1 11 78 19
CC-A3* 3 7 2 2 2 13
CC-A4 11
CC-A5 1
CC-A6 2
CC-A7* 1 2 1
CC-A8 1 1
CC-A9* 1 1
CC-A10* 2 3 5 1
CC-A11 1
CC-A12* 1 1
CC-A13 2
CC-A14 3
CC-A15 1
CC-A16 1
CC-A17 1
CC-A19* 1
Totals 45 131 42 50 105 20 11 81 32
butions to the Chiriqu? Lagoon foraging ground 
were divided between the South Florida (0.658 to 0.712)
and Mexico (0.281 to 0.342) populations, while no other
nesting beaches make significant contributions. If the
Mediterranean stocks are included among possible
sources (right columns of Table 2), the estimated con-
tribution of these stocks is uniformly high (0.367?0.436
for Greece?SE Med.; 0.095?0.103 for Turkey), the con-
tribution of Mexico remains similar (0.171 to 0.208), and
the contribution from the Florida panhandle population,
which consists of only a few hundred nesting females,
is also very high (0.233?0.292). In these analyses the
estimated contribution from the largest nesting popula-
tion in the Atlantic, South Florida, is minimal
(0.0?0.116). Contributions of Atlantic nesting beach
populations estimated by each of the three mixed stock
analysis programmes differ significantly from expected
nesting beach contributions based on numbers of nests
recorded at Atlantic and Mediterranean nesting beaches
(Table 3).
DISCUSSION
Mixed stock analysis
On general principles one would hope that the best
mixed stock analysis result would proceed from the
inclusion of all available genetic data from all possible
sources. In this case, with the genetic markers and
analytical tools we have available, we believe that this
simple approach is not the most effective and that it is
reasonable to use additional non-genetic data to rule out
unrealistic maximum likelihood solutions. Although it is
possible that some of the turtles in Chiriqu? Lagoon are
from Mediterranean beaches, the inclusion of these
beaches as possible sources leads to the conclusion that
over half of the turtles in Chiriqu? Lagoon are from the
Mediterranean and another third are from the tiny
Florida panhandle population. If this were true it would
represent an extraordinary biological phenomenon. The
only record of recruitment of juvenile turtles from
Mediterranean stocks to developmental or pelagic
habitats in the Atlantic involves three animals out of a
sample of 300 from developmental habitats in east
Florida (D. Bagley, unpublished data). No other study
has reported any contribution of Mediterranean stocks
(Sears et al., 1995; Bass et al., 1998; Bolten et al., 1998;
Lahanas et al., 1998: Bass & Witzell, 2000; Rankin-
Baransky et al., in press). It would also be difficult to
propose a reasonable biological mechanism to explain
the 20?30% contribution of the Florida panhandle
population, whereas the neighbouring South Florida
population, which is several orders of magnitude larger,
contributes nothing. 
Although the biological improbability of these results
does not preclude their veracity, such an extraordinary
result does encourage a closer examination of the method-
ology for possible sources of bias. The maximum likeli-
hood algorithms employed here are known to produce
spurious results under certain conditions: (1) if the ratio
of observed haplotypes to populations examined is less
than 1.5 (Xu et al., 1994); (2) if haplotype frequencies
129Origin of juvenile loggerheads in Panam?
Table 2. Estimated nesting beach contributions to the Chiriqu? Lagoon, Panam?, developmental habitat based on three maximum likelihood
mixed stock analysis programmes. Results for analyses including only Atlantic nesting beaches as possible contributing sources are shown in
the left-hand columns. Results for analyses including both Atlantic and Mediterranean nesting beaches are shown on the right.
Population CONSQRT contribution (SD) UCON contribution SHADRACQ contribution
Atlantic only Atl. & Med. Atlantic only Atl. & Med Atlantic Only Atl. & Med.
South Florida 0.6814 (0.1326) 0.0172 (0.0199) 0.7119 0.1162 0.6580 0.0000
Mexico 0.3181 (0.1326) 0.2079 (0.1162) 0.2811 0.1775 0.3420 0.1712
Northern (NEFL?NC) 0.0000 (0.0000) 0.0000 (0.0000) 0.0002 0.0028 0.0000 0.0000
Florida panhandle 0.0006 (0.0006) 0.2896 (0.0784) 0.0068 0.2331 0.0000 0.2921
Brazil 0.0000 (0.0000) 0.0000 (0.0000) 0.0000 0.0000 0.0000 0.0000
Greece/SE Med NA 0.3905 (0.1314) NA 0.3669 NA 0.4357
Turkey NA 0.0948 (0.1087) NA 0.1035 NA 0.1009
Table 3. Comparison between maximum likelihood estimates of stock contributions and expected contributions of nesting beaches based on
proportions of nesting females. Population size estimates for US beaches are based on data compiled by the Turtle Expert Working Group
(TEWG, 2000) and represent mean number of females nesting annually between 1989 and 1995. Population estimates for Mexico are from
Zurita, Herrera & Prezas (1993) and for Brazil are from Marcovaldi & Laurent (1996). Contribution proportions from Table 2 are converted
into number of turtles expected in our sample of 43. 
Nesting Population Expected contribution Maximum likelihood estimates
population estimate CONSQRT UCON SHADRACQ
Females
annually Proportion # Turtles % # % # % #
South Florida 15,595 85.75% 36.87 68.140 29.30 71.19 30.61 65.80 28.29
Mexico 561 3.08% 1.33 31.810 13.68 28.11 12.09 34.20 14.71
Northern (NEFL-NC) 1519 8.35% 3.59 0.000 0 0.02 0.01 0.00 0
Florida Panhandle 111 0.61% 0.26 0.060 0.02 0.68 0.29 0.00 0
Brazil 400 2.20% 0.95 0.000 0 0.00 0 0.00 0
?2 ? 110.06 ?2 ? 83.75 ?2 ? 129.60
P < 0.001 P < 0.001 P < 0.001
differ by less than 30% among populations (Broderick et
al., 1994); (3) if genetic diversity of source populations
differs greatly and the haplotypes in the less diverse pop-
ulations are a subset of those found in the more diverse
populations (Davies, Villablanca & Roderick, 1999). All
three of these conditions exist in the data set that includes
the Mediterranean beaches. Furthermore, haplotype CC-
A6 (F), which is the only endemic Mediterranean haplo-
type, was not detected in our sample, so all contributions
ascribed to the Mediterranean could also be attributed to
Atlantic populations. 
We suggest that the exclusion of some source popu-
lations from mixed stock analyses is justified if it is bio-
logically the most reasonable alternative and if the
inclusion of those populations produces analytical con-
ditions that are known to bias results. We feel that both
biological and analytical lines of evidence strongly indi-
cate that analyses that exclude the Mediterranean
beaches are more accurate. Thus we have chosen to
focus on results of analyses that exclude these popula-
tions. The development of molecular markers that unam-
biguously distinguish the Mediterranean from the
Atlantic nesting populations will be an important avenue
for future studies.
Patterns of juvenile loggerhead recruitment
Lahanas et al. (1998) observed that dispersal during the
pelagic stage provides the potential thoroughly to mix
turtles from separate nesting populations and that the
content of this pelagic melting pot may be reflected in
subsequent recruitment to benthic developmental habi-
tats. Under this turtle soup model, the composition of
the pool of post-hatchling turtles in the pelagic stage
would be proportional to the sizes of contributing stocks.
If subsequent recruitment of juvenile turtles to benthic
developmental habitats represented an unbiased spoon-
ful from this pool of potential recruits, then the propor-
tions of juvenile turtles in benthic developmental
habitats would also be expected to reflect the sizes of
contributing stocks. Studies of pelagic-stage loggerhead
turtles (Bolten et al., 1998) in the eastern Atlantic and
green turtles (Chelonia mydas) in developmental habi-
tat in the Bahamas (Lahanas et al., 1998) are consistent
with this turtle soup model. Alternatively, any number
of oceanographic and behavioural factors, including
proximity to nesting beaches (Lahanas et al., 1998), local
current anomalies (Carr & Meylan, 1980) and selective
settlement or directed migration, could lead to dispro-
portionate nesting beach contributions to developmental
habitats even if contributions to the pelagic habitats are
proportional to rookery size.
Evidence both from direct comparisons of haplotype
frequencies in Chiriqu? Lagoon with haplotype fre-
quencies in the pelagic habitat and from mixed stock
analysis indicates that recruitment of juvenile logger-
heads to Chiriqu? Lagoon is not consistent with the
simple, turtle soup model. Bolten et al. (1998) found that
nesting beach contributions to pelagic loggerhead habitat
in the eastern Atlantic were proportional to the sizes of
contributing rookeries. Haplotype frequencies in the
Chiriqu? Lagoon developmental habitat differ signifi-
cantly from the haplotype frequencies in the eastern
Atlantic pelagic sample, indicating that loggerhead
turtles in Chiriqu? Lagoon do not represent a random
selection from the pool of potential recruits in the east-
ern Atlantic pelagic habitat. Similarly, mixed stock
analysis estimates of the Chiriqu? sample indicate that
approximately 70% of the turtles in Chiriqu? Lagoon
originate from South Florida and 30% from Mexico. If
recruitment to Chiriqu? Lagoon were proportional to the
sizes of nesting beaches in the Atlantic, approximately
86% of the loggerhead turtles captured in Chiriqu?
Lagoon would originate from South Florida, 8% from
the Northern population (northeast Florida?North
Carolina), 3% from Mexico and 2% from Brazil. The
overrepresentation of Mexican turtles in our sample indi-
cates that recruitment to this developmental habitat is
not random and is likely driven by oceanographic or
behavioural factors other than simply the size of
contributing populations. Evidence is slowly accumulat-
ing for the predominance of this non-random recruitment
in marine turtles (Sears et al., 1995; Laurent et al., 1998;
Bass & Witzell, 2000; Rankin-Baransky et al., in press;
D. Bagley, unpublished data). As a caveat, it is impor-
tant to realize that although we have used both nesting
beach population estimates and maximum likelihood
estimates in quantitative analyses, there are inherent lim-
itations of these data for such purposes. The calculations
of expected nesting beach contributions are based on
population estimates with unknown error. The mixed
stock analysis is hampered by the small sample size of
43 individuals available for this study and by incomplete
nesting beach sampling. Given these limitations, it is
important that even the quantitative results presented
here be interpreted in a qualitative manner. 
Conservation implications
Chiriqu? Lagoon supports a significant population of
immature loggerheads. More than one-third (81 of 216)
of the turtles captured in Chiriqu? Lagoon by the Bocas
Turtle Project have been loggerheads. The presence of
significant numbers of Mexican and US turtles in
Chiriqu? Lagoon shows that this area, and perhaps other,
as-yet-undescribed feeding grounds in the Caribbean, are
important developmental habitats for these nesting
populations. Mortality sources in Caribbean develop-
mental habitats are not currently included in demo-
graphic models for Florida or Mexican loggerhead
populations (TEWG, 2000). Mortality sources in
Caribbean developmental habitats could be particularly
detrimental to the persistence of the Mexican loggerhead
population because it is small (~500 females) and there-
fore demographically vulnerable. Demographic models
for marine turtles indicate that population growth is
sensitive to mortality of the subadult size class that 
is present in Chiriqu? Lagoon (Crouse, Crowder &
Caswell, 1987) and that harvest of just a few hundred
subadult individuals from a small population can lead to
130 T. N. ENGSTROM ET AL.
population decline (Heppell et al., 1996). Tag return data
indicate that at least seven of the 81 (8.6%) loggerhead
turtles tagged in Chiriqu? Lagoon since 1987 have been
subsequently captured in subsistence fisheries in
Chiriqu? Lagoon and elsewhere in the Caribbean. Thus,
the persistence of the Mexican loggerhead population
may be threatened because subadult turtles from this
population are concentrated in Chiriqu? Lagoon, and per-
haps in other developmental habitats in the Caribbean
where they are subject to harvest in small-scale fisheries.
Although the Mexican nesting population is small, it is
the most genetically diverse loggerhead population in
the Atlantic or Mediterranean (Encalada, Zurita &
Bowen, 1999). The loss of this nesting population would
result in the loss of 30% of the mtDNA diversity in the
Atlantic basin (Encalada et al., 1999). 
The presence of two new haplotypes in a sample of
just 45 individuals from Chiriqu? Lagoon indicates that
genetic diversity on loggerhead nesting beaches is not
completely known. The novel haplotypes observed in
Chiriqu? Lagoon may be present at a low frequency in
the populations surveyed (Encalada et al., 1998) and
may have simply escaped detection because of sampling
error. Alternatively, they may represent markers present
in other nesting populations that have not yet been
surveyed. In addition to the major nesting beaches in the
US, Mexico and Brazil, loggerhead nesting occurs in
several areas throughout the Atlantic (Fig. 1), including
Colombia, the Caribbean coast of Central America,
Cuba, Jamaica, the Cayman Islands, the Dominican
Republic, Haiti, Puerto Rico and the Bahamas
(Kaufman, 1975; Carr et al., 1982; Ehrhart, 1989;
Addison & Morford, 1996; Amorocho, C?rdoba &
Milkin, 1999; Aiken et al., 2001). It is possible that these
other small nesting populations are also contributing to
Caribbean feeding grounds and would be similarly
adversely affected by fisheries in the region. It is par-
ticularly interesting to consider the possibility that a
Caribbean pool of advanced juveniles may persist from
the once substantial Colombian rookery, which has col-
lapsed in the last 20 years because of overharvesting
(Amorocho et al., 1999). Expanding the current nesting
beach sampling to include Colombia and other small
Caribbean beaches will enable us to identify the origin
of turtles bearing these unknown haplotypes and to
determine what, if any, contribution the small, nesting
populations make to Chiriqu? Lagoon and other
Caribbean feeding grounds. 
Based on the results of this study, we are able to make
several recommendations regarding future research and
conservation efforts specific to Chiriqu? Lagoon and for
marine turtle conservation in general.
Consideration should be given to extending the exist-
ing Bastimentos Island National Marine Park (Fig. 1) to
include the eastern part of the Chiriqu? Lagoon in which
important marine turtle developmental habitat is found.
By doing so, the government of Panam? would be grant-
ing protection to loggerhead turtles from Florida, from
the vulnerable Mexican nesting population, and perhaps
from other small, Caribbean nesting populations.
The documentation of incidental and direct take and
the collection of genetic samples for mixed stock analy-
sis should be a standard part of all studies of marine tur-
tles on feeding grounds. These two types of data are
necessary to access accurately the possible effects of har-
vest and incidental take in an area. In this study, we have
shown that a disproportionately large number of
Mexican loggerhead turtles inhabit Chiriqu? Lagoon and
have suggested that the local artisanal fishery may com-
promise the Mexican population. Although some data
have been collected on this fishery (R. Landrum &
P. Meylan, unpublished data), no attempt has been made
to describe quantitatively the level of harvest. This is an
important area of future research in Chiriqu? Lagoon. It
is essential that these data on mortality in Caribbean
habitats be included in demographic models and man-
agement plans of both Mexican and Florida populations. 
An effort should be made to increase sampling at
surveyed beaches and to get samples from smaller,
unsurveyed beaches, particularly in Colombia, in order
to detect rare haplotypes. A concomitant effort should
be made to identify and genetically characterize other
tropical feeding grounds. The combination of these data
would allow us to describe more accurately the geo-
graphic distribution of feeding grounds used by each
nesting population and thereby predict more accurately
the effects of mortality on those feeding grounds. 
Finally, our results reinforce once again that marine
turtles, when viewed as a resource, must be considered
to be an internationally shared resource. Management of
the western Atlantic marine turtle populations requires
cooperation from countries spanning large political,
environmental and geographic spectra, because these
animals respect neither political nor climatological
boundaries. Turtles from temperate nesting beaches in
the US and from the more southerly populations from
Mexico mix in tropical developmental habitat in
Panam?. It is clear that marine turtle conservation efforts
must focus not only on protection of nesting habitat but
also on identification and protection of turtles in pelagic,
developmental and adult foraging habitats. 
Acknowledgements
This research was carried out with permission from the
Panamanian Protected Areas and Wildlife, National
Authority for the Environment (ANAM). Blood samples
were imported to the US under CITES permit
US758093. Major funding was provided by a grant to
P. & A. Meylan from the Wildlife Conservation Society,
with additional funding or support from Eckerd College,
the Ford Foundation, Florida Fish and Wildlife
Conservation Commission, and an NSF grant to H. B.
Shaffer. Invaluable logistical and technical support was
provided by the Smithsonian Tropical Research Institute.
The Chiriqu? Land Company provided transportation on
numerous occasions and contributed sterilized drums 
for our water supply. Allen Foley, Jim Quinn, Blair
Whitherington and an anonymous reviewer provided
helpful comments on the manuscript. We thank
131Origin of juvenile loggerheads in Panam?
D. Bagley, Anna Bass and Brian Bowen for use of un-
published information and the ACCSTR for information
on haplotype designation. R. S. Duncan, R. Landrum,
L. Castillo, G. Castillo and Poppa D. provided assistance
in the field. Chencho Castillo was an indispensable
source of turtle knowledge, history, humour and myths.
REFERENCES
Addison, D. S. & Morford, B. (1996). Sea turtle nesting activity
on the Cay Sal Bank, Bahamas. Bahamas J. Sci. 3: 31?36.
Aiken, J. J., Godley, B. J., Broderick, A., Austin, T., Ebanks-
Petrie, G. & Hays, G. C. (2001). Two hundred years after a com-
mercial marine turtle fishery: the current status of marine turtles
nesting in the Cayman Islands. Oryx 35(2):145?151. 
Amorocho, D., C?rdoba, B. J. A. & Milkin, H. S. (1999). Current
status of nesting sea turtles in northern Colombian Caribbean.
Mar. Turt. Newslet 85: 6?7.
Bass, A. L., Lagueux, C. J. & Bowen, B. W. (1998). Origin of
green turtles, Chelonia mydas, at ?sleeping rocks? off the north-
east coast of Nicaragua. Copeia 1998: 1064?1069.
Bass, A. L. & Witzell, W. N. (2000). Demographic composition
of immature green turtles (Chelonia mydas) from the east cen-
tral Florida coast: evidence from mtDNA markers.
Herpetologica 56: 357?367.
Bolten, A. B. (1999). Techniques for measuring sea turtles. In
Research and management techniques for the conservation of
sea turtles: 110?114. Ekert, K. L., Bjorndal, K. A., Abreu-
Grobois, F. A. & Donnelly, M. (Eds). IUCN/SSC Marine Turtle
Specialist Group Publication No. 4.
Bolten, A. B. & Balazs, G. H. (1995). Biology of the early pelagic
stage: the ?lost year?. In Biology and conservation of sea tur-
tles: 579?581. Bjorndal, K. A. (Ed.). Washington, DC:
Smithsonian Institution Press.
Bolten, A. B., Bjorndal, K. A., Martins, H. R., Dellinger, T.,
Biscoito, M. J., Encalada, S. E. & Bowen, B. W. (1998).
Transatlantic developmental migrations of loggerhead turtles
demonstrated by mtDNA sequence analysis. Ecol. Applic. 8:
1?7.
Bowen, B. W. (1995). Molecular genetic studies of marine tur-
tles. In Biology and conservation of sea turtles: 585?587.
Bjorndal, K. A. (Ed.). Washington, DC: Smithsonian Institution
Press.
Bowen, B. W., Abreu-Grobois, F. A., Balazs, G. H., Kamezaki,
N., Limpus, C. J. & Ferl, R. J. (1995). Trans-Pacific migrations
of the loggerhead turtle (Caretta caretta) demonstrated with
mitochondrial DNA markers. Proc. Natl. Acad. Sci. USA. 92:
3731?3734.
Bowen, B. W., Bass, A. L., Garcia-Rodriguez, A., Diez, C. E.,
van Dam, R., Bolton, A., Bjorndal, K. A., Miyamoto, M. M. &
Ferl, R. J. (1996). Origin of hawksbill turtles in a Caribbean
feeding area as indicated by genetic markers. Ecol. Applic. 6:
566?572.
Bowen, B. W. & Karl, S. A. (1997). Population genetics, phylo-
geography, and molecular evolution. In The biology of sea tur-
tles: 29?50. Lutz, P. & Musick, J. (Eds). Boca Raton, FL: CRC
Press, Inc.
Broderick, D., Moritz, C., Miller, J. D., Guinae, M., Prince,
R. I. T. & Limpus, C. J. (1994). Genetic studies of the hawks-
bill turtle Eretmochelys imbricata: evidence for multiple stocks
in Australian waters. Pacific Conserv. Biol. 1: 123?132.
Carr, A. F. (1986). Rips, FADS and little loggerheads. BioScience
36: 92?100.
Carr, A. F. (1987). New perspectives on the pelagic stage of sea
turtle development. Conserv. Biol. 1: 103?121.
Carr, A. F., Carr, M. H. & Meylan, A. B. (1978). Ecology and
migrations of sea turtles, 7: The west Caribbean green turtle
colony. Bull. Am. Mus. Nat. Hist. 162: 1?46.
Carr, A. F. & Meylan, A. B. (1980). Evidence of passive migra-
tion of green turtle hatchlings in Sargassum. Copeia 1980:
366?368.
Carr, A. F., Meylan, A. B., Mortimer, J. A., Bjorndal, K. A. &
Carr, T. (1982). Surveys of sea turtle populations and habitats
in the Western Atlantic. US Department of Commerce, Panama
City, FL.
Crouse, D. A., Crowder, L. B. & Caswell, H. (1987). A stage-
based population model for loggerhead turtles and implications
for conservation. Ecology 68: 1412?1423.
Davies, N., Villablanca, F. X. & Roderick, G. K. (1999).
Determining the source of individuals: multilocus genotyping in
nonequilibrium population genetics. TREE 14: 17?21.
Ehrhart, L. (1989). Status report of the loggerhead turtle. In
Proceedings of the Second Western Atlantic Turtle Symposium:
122?139. Ogren, L. et al. (Eds). NOAA Technical Memorandum
NMFS-SEFC-226.
Encalada, S. E., Bjorndal, K. A., Bolten, A. B., Zurita, J. C.,
Schroeder, B., Possardt, E., Sears, C. J. & Bowen, B. W. (1998).
Population structure of loggerhead turtle (Caretta caretta) nest-
ing colonies in the Atlantic and Mediterranean as inferred from
mitochondrial DNA control region sequences. Mar. Biol. 130:
567?575.
Encalada, S. E., Zurita, J. C. & Bowen, B. W. (1999). Genetic
consequences of coastal development: the sea turtle rookeries at
X?cacel, Mexico. Mar. Turt. Newslet, 83: 8?10.
Heppell, S. S., Limpus, C. J., Crouse, D. T., Frazer, N. B. &
Crowder, L. B. (1996). Population model analysis for the log-
gerhead sea turtle, Caretta caretta, in Queensland. Wildl. Res.
23: 143?159.
Kaufman, R. (1975). Studies of the loggerhead sea turtle, Caretta
caretta caretta (Linn?) in Colombia, South America.
Herpetologica. 31: 323?326.
Lahanas, P. N., Bjorndal, K. A., Bolten, A. B., Encalada, S. E.,
Miyamoto, M. M., Valverde, R. A. & Bowen, B. W. (1998).
Genetic composition of a green turtle (Chelonia mydas) feeding
ground population: evidence for multiple origins. Mar. Biol.
130: 345?352.
Laurent, L., Casale, P., Bradai, M. N., Godley, B. J., Gerosa, G.,
Broderick, A. C., Schroth, W., Schierwater, B., Levy, A. M.,
Freggi, D., Abd El-Mawla, E. M., Hadoud, D. A., Gomati,
H. E., Domingo, M., Hadjichristophorou, M., Kornaraky, L.,
Demirayak, L. F. & Gautier, C. (1998). Molecular resolution of
marine turtle stock composition in fishery bycatch: a case study
in the Mediterranean. Mol. Ecol. 7: 1529?1542.
Limpus, C. J., Couper, P. J. & Read, M. A. (1994a). The green
turtle, Chelonia mydas, in Queensland: population structure in
a warm temperate feeding area. Mem. Queensland Mus. 35:
139?154.
Limpus, C. J., Couper, P. J. & Read, M. A. (1994b). The log-
gerhead turtle, Caretta caretta, in Queensland: population struc-
ture in a warm temperate feeding area. Mem. Queensland Mus.
37: 195?204.
Limpus, C. J., Miller, J. D., Parmenter, C. J., Reimer, D.,
McLachlan, N. & Webb, R. (1992). Migration of green
(Chelonia mydas) and loggerhead (Caretta caretta) turtles to and
from eastern Australian rookeries. Wildl. Res. 19: 347?358.
Limpus, C. J. & Reed, P. C. (1985). The loggerhead turtle, Caretta
caretta, in Queensland [Australia]: observations on internesting
behavior. Austral. Wildl. Res. 12: 535?540.
Marcovaldi, M. A. & Laurent, A. (1996). A six season study of
marine turtle nesting at Praia do Forte, Bahia, Brazil, with impli-
cations for conservation and management. Chelonian Conserv.
Biol. 2: 55?59.
Masuda, M., Nelson, S. & Pella, J. (1991). User manual for
GIRSEM, GIRLSYM and CONSQRT. USA-DOC-NOAA-NMFS.
US-Canada Salmon Program, Juneau, Alaska.
Meylan, A. B. (1982). Sea turtle migration: evidence from tag
returns. In Biology and conservation of sea turtles: 91?100.
132 T. N. ENGSTROM ET AL.
Bjorndal, K. A. (Ed.). Washington, DC: Smithsonian Institution
Press.
Meylan, A. B., Bowen, B. W. & Avise, J. C. (1990). A genetic
test of the natal homing versus social facilitation models for
green turtle migration. Science 248: 724?727.
Norman, J. A., Moritz, C. & Limpus, C. J. (1994). Mitochondrial
DNA control region polymorphisms: genetic markers for eco-
logical studies of marine turtles. Mol. Ecol. 3: 363?373.
Owens, D. W. & Riuz, G. J. (1980). New methods of obtaining
blood and cerebrospinal fluids from marine turtles.
Herpetologica 36: 17?20.
Palumbi, S. R. (1996). Nucleic acids II: the polymerase chain
reaction. In Molecular systematics, second edition: 205?247.
Hillis, D. M., Maritz, C. & Mable, B. K. (Eds). Sunderland, MA:
Sinauer Associates.
Pella, J. J. & Milner, G. B. (1987). Use of genetic marks in stock
composition analysis. In Population genetics and fishery man-
agement: 247?276. Ryman, N. & Utter, F. (Eds). Seattle:
University of Washington Press.
Rankin-Baransky, K., Williams, C. J., Bass, A. L., Bowen, B. W.
& Spotila, J. R. (In press). Origin of loggerhead turtles stranded
in the northeast United States as determined by mt DNA analy-
sis. J. Herpetol.
Saiki, R. K., Bugawan, T. L., Horn, G. T., Mullis, K. B. & Erlich,
H. A. (1986). Analysis of enzymatically amplified beta globin
and Hla-Dq-alpha DNA with allele-specific oligonucleotide
probes. Nature 324: 163?166.
Sears, C. J., Bowen, B. W., Chapman, R. W., Galloway, S. B.,
Hopkins-Murphy, S. R. & Woodley, C. M. (1995). Demographic
composition of the feeding population of juvenile loggerhead
turtles (Caretta caretta) off Charleston, South Carolina: evi-
dence from mitochondrial DNA markers. Mar. Biol. 123:
869?874.
Sokal, R. R. & Rohlf, F. J. (1995). Biometry: the principles and
practice of statistics in biological research. New York: W. H.
Freeman.
TEWG (Turtle Expert Working Group) (2000). Assessment update
for the Kemp?s Ridley and loggerhead sea turtle populations in
the western North Atlantic. US Department of Commerce,
NOAA Technical Memorandum NMFS-SEFCS-444.
Xu, S., Kobak, C. J. & Smouse, P. E. (1994). Constrained least
squares estimation of mixed population stock composition from
mtDNA haplotype frequency data. Can. J. Fish. Aq. Sci. 51:
417?425.
Zurita, J. C., Herrera, R. & Prezas, B. (1993). Biolog?a y conser-
vaci?n de las tortugas marinas en el litoral de Quintana Roo,
Temporada (1990). In IV Taller regional sobre programas de
conservaci?n de tortugas marinas en la pen?nsula de Yucat?n.
Frazier, J. (Ed.). Mexico: Universidad Aut?noma de Yucat?n,
M?rida, Yucat?n, M?xico.
133Origin of juvenile loggerheads in Panam?