Mitochondrial and nuclear DNA analyses reveal fine scale geographic structure in bottlenose dolphins ( Tursiops truncatus ) in the Gulf of Mexico

Ó Springer 2005 Conservation Genetics (2005) 6:715–728 DOI 10.1007/s10592-005-9031-7 Mitochondrial and nuclear DNA analyses reveal fine scale geographic structure in bottlenose dolphins (Tursiops truncatus) in the Gulf of Mexico Anna B. Sellas1,3, Randall S. Wells2 & Patricia E. Rosel3,* 1 Ocean Sciences Department, University of California, Santa Cruz, CA, 95064, USA; 2Chicago Zoological Society, c/o Mote Marine Laboratory, 1600 Ken Thompson Parkway, Sarasota, FL, 34236, USA; 3National Marine Fisheries Service, 646 Cajundome Blvd, Lafayette, LA, 70506, USA (*Corresponding author: Phone: +1-337-291-2123; Fax: +1-337-291-2106; E-mail: Patricia.Rosel@noaa.gov) Received 30 September 2004; accepted 10 February 2005 Key words: management, microsatellites, mitochondrial DNA, population structure, Tursiops truncatus Abstract There is a need for biological information to support current stock designations of bottlenose dolphins (Tursiops truncatus) in the Gulf of Mexico. The existence of many inshore, resident ‘‘communities’’ raises questions as to the relationship these dolphins may hold with dolphins inhabiting neighboring inshore and coastal areas. In this study, population subdivision was examined among four resident, inshore bottlenose dolphin stocks (Sarasota Bay, FL, Tampa Bay, FL, Charlotte Harbor, FL and Matagorda Bay, TX) and one coastal stock (1–12 km offshore) in the Gulf of Mexico. Evidence of significant population structure among all areas was found on the basis of both mitochondrial DNA (mtDNA) control region sequence data and nine nuclear microsatellite loci. Estimates of relatedness showed no population contained a significantly high number of related individuals, while separate AMOVAs for males and females indicated that both sexes exhibit a significant level of site philopatry. Results presented here provide the first genetic evidence of population subdivision between the coastal Gulf of Mexico and adjacent inshore areas along the central west coast of Florida. Such strong genetic subdivision is surprising given the short geographical distance between many of these areas and the lack of obvious geographic barriers to prevent gene flow. These findings support the current, separate identification of stocks for bottlenose dolphins inhabiting the eastern coastal and inshore areas of the Gulf of Mexico. Introduction For cetacean species, Hoelzel (1998) suggested an important mechanism for the formation of intraspecific genetic differentiation within a geographic region to be behavior specialization which, in turn, may be the precursor to resource specialization (Skúlason and Smith 1995). One example of intraspecific differentiation within a geographic region in cetacean species is found in bottlenose dolphins (Tursiops truncatus) where strong differ- ences have been documented between coastal and offshore forms in both the Northwest Atlantic and Gulf of Mexico on the basis of hematology, morphology, stomach contents, parasite loads and genetics (Hersh and Duffield 1990; Ross and Cockcroft 1990; Banos and Odell 1990; Mead and Potter 1995; Hoelzel et al. 1998; Kingston and Rosel 2003). While the level of differentiation between coastal and offshore Tursiops is well documented, and may eventually lead to separate taxonomic species designations, the distinctions 716 among inshore, and between inshore and coastal areas, are less well studied. Here we define inshore areas as bays, sounds or estuarine waters, bordered by mainland on one side and barrier islands on the other, and coastal waters as nearshore or neritic waters from shore to 12 km seaward of shore. From long-term photo-identification and telemetry studies, it appears many inshore areas of both the Gulf of Mexico and the Northwest Atlantic harbor resident bottlenose dolphins that generally remain in a discrete geographic area over time (Irvine and Wells 1972; Shane et al. 1986; Wells 1986; Würsig and Lynn 1996; Zolman 2002; Wells 2003). From a management perspective, the preservation of each of these resident stocks is important since repopulation of any area would theoretically require a recolonization event. Therefore, in the Gulf of Mexico, each of 33 inshore areas is currently managed as a separate stock (Waring et al. 2002). Along the eastern coast of the Gulf of Mexico, running adjacent to many inshore areas, a separate coastal stock has been defined as ranging from shore to 9.3 km seaward of the 18.3 m isobath (Waring et al. 2002). Two additional bottlenose dolphin stocks range farther offshore in the Gulf of Mexico (Waring et al. 2002) but will not be considered in this paper. Due to the many logistical obstacles involved in studying wild cetaceans, biological data to support the discreteness of most of these bottlenose dolphin stocks is lacking. Bottlenose dolphins inhabiting inshore waters along the central west coast of Florida have been the subjects of research since 1970 (Irvine and Wells 1972; Wells et al. 1980; Scott et al. 1990; Wells 1991; 2003). While these research efforts have focused on the resident dolphins in Sarasota Bay, FL, considerable research has also been conducted in the two adjacent inshore areas, Tampa Bay, FL, to the north, and Charlotte Harbor, FL (including Pine Island Sound), to the south (Wells et al. 1996a; b; Shane et al. 1986). In addition to these inshore areas, photo-identification data have also been collected in the neighboring coastal area of the Gulf of Mexico from Tampa Bay down to Lemon Bay, FL (Fazioli 1999) and, in the western Gulf of Mexico, in Matagorda Bay, TX (Würsig and Lynn 1996). These studies have all documented long-term year-round use of each of these areas by core groups of resident dolphins. Nevertheless, for the inshore Florida localities, interactions of dolphins between neighboring areas are not uncommon. Wells (1986) estimated 17% of group sightings of resident Sarasota Bay dolphins also included at least one non-resident. In the adjacent coastal Gulf of Mexico waters, Fazioli (1999) found 14% of group sightings of coastal Gulf dolphins also involved at least one resident Sarasota Bay dolphin. Similarly, mixed groups containing Sarasota Bay and Tampa Bay dolphins, and mixed groups containing Tampa Bay and coastal Gulf dolphins, are also commonly observed (Weigle 1990; Wells 1991). These mixed groups suggest a certain amount of genetic exchange could occur among dolphins inhabiting different inshore areas and/or between dolphins in inshore and coastal areas. Surprisingly few studies have investigated the degree of genetic exchange that occurs between these well documented inshore dolphin stocks. From the few studies available, it has been suggested that, while resident, inshore bottlenose dolphins often remain in their local area over time, gene flow between adjacent, inshore areas remains high enough to maintain homogeneity (Dowling and Brown 1993; Duffield and Wells 2002). Using whole mtDNA restriction fragment length (RFLP) analysis and six endonucleases, Dowling and Brown (1993) did not detect significant subdivision among samples from the Gulf of Mexico, although samples sizes were small in some regions. Duffield and Wells (1991, 2002) found allele frequency differences at three allozyme loci between Sarasota Bay and both Tampa Bay and Charlotte Harbor, but did not detect differences between the latter two bays. A similar pattern in haplotype frequency differences was seen using a HinfI digest of whole mtDNA. (Duffield and Wells 1991, 2002). Consequently, questions remain as to the level of genetic differentiation that exists between coastal and inshore areas and, on a finer scale, among inshore areas. Indeed the extreme mobility of these dolphins, coupled with a lack of obvious geographic barriers and field data documenting long-range movements and mixing of dolphins from the inshore and coastal areas, would imply that there is little reproductive isolation between them. While this is a logical conclusion, mixing of animals does not necessarily mean that interbreeding is occurring. This study uses 717 genetic methods to measure, indirectly, the level of gene flow among four well documented inshore stocks (Tampa Bay, FL, Sarasota Bay, FL, Charlotte Harbor, FL, and Matagorda Bay, TX) and one putative coastal stock of bottlenose dolphins in the Gulf of Mexico. Mitochondrial DNA control region sequences and nuclear microsatellites were used to test the null hypothesis that resident bottlenose dolphins of these inshore areas and coastal bottlenose dolphins in the Gulf of Mexico represent one homogeneous population. Moreover, if significant genetic differences do exist between these proximal areas, what mechanism(s), such as behavior and/or resource specialization, might explain this differentiation? Methods Tampa Bay, Charlotte Harbor, and the coastal Gulf of Mexico. In Sarasota Bay and Matagorda Bay, skin samples were collected during live capture-release projects using a surgical biopsy procedure. The coastal Gulf study area extended approximately 12 km offshore from just outside of Tampa Bay (27.62° N–82.91° W) to the south end of Lemon Bay (26.84° N–82.49° W), a distance of approximately 94 km (Figure 1). Genomic DNA was extracted from the samples using standard proteinase K digestion and phenol/chloroform extraction protocols, as outlined in Rosel and Block (1996). Samples collected via biopsy darting were sexed using ZFX and SRY specific primers as described in Rosel (2003). In Sarasota Bay, where extensive genealogy data were available for many dolphins, when samples were available from mothers and their calves, only the sample from the mother was included in the analyses. Sample collection MtDNA sequencing Skin samples from 223 dolphins were collected between 1996 and 2002 (Figure 1). Skin samples were collected using a biopsy dart method in A 450 bp fragment, including the proline tRNA and the 5¢ end of the mitochondrial DNA control Figure 1. Map showing locations where samples were collected. Sampling areas are depicted by either solid black circles (inshore – Tampa Bay, Sarasota Bay, Charlotte Harbor, Matagorda Bay) or a hashed box (coastal – Gulf of Mexico). 718 region, was amplified via PCR using primers L15824 (5¢- CCTCACTCCTCCCTAAGACT-3¢) and H16265 (5¢-GCCCGGTGCGAGAAGAGG3¢) (Rosel et al. 1999b). Amplification reactions contained 50 ng DNA, 20 mM Tris–HCl, pH 8.0, 50 mM KCl, 1.5 mM MgC12, 0.3 lM of each primer, 150 lM dNTPs and 1.25 units of Taq DNA polymerase. The thermal cycler profile consisted of initial denaturation at 94 °C for 30 s, followed by 30 cycles of 94 °C for 1 min, 52 °C for 1 min and 72 °C for 1 min, with a final extension at 72 °C for 10 min. PCR products were purified by excision from a 0.8% SeaPlaqueÒ GTGÒ low melting point agarose gel (FMC, Rockland, ME) and the agarose bands digested overnight at 40 °C with 5–10 units of agarase (Sigma, St. Louis, MO). Cycle sequencing reactions were carried out using the original forward and reverse primers with 3– 6 ll of the agarased product following general protocols supplied from the manufacturer for ABI Prism Big Dye TerminatorÒ Ready-reaction mix (Applied Biosystems, Foster City, CA). Cycle sequencing products were purified either by ethanol precipitation or using Centri-Sep spin columns (Princeton Separations, Adelphia, NJ) and then sequenced using an ABI 310 Genetic Analyzer. Sequences were edited using the program Sequence Navigator 1.0.1 (Applied Biosystems, Foster City, CA). Consensus sequences were aligned by eye using the program SeqPup 0.6 (Gilbert 1995) and used for mtDNA analyses. MtDNA sequence analysis Nucleotide (p) and haplotypic (d) diversities (Nei 1987) were estimated for each stock using Arlequin 2.0 (Schneider et al. 2000). To measure the degree of population subdivision, an analysis of molecular variance (AMOVA, Excoffier et al. 1992) was run using Arlequin 2.0. The AMOVA calculates a measure of population subdivision based on haplotype frequency data alone, FST (Wright 1965), and a measure, FST, based on both haplotype frequency and genetic distance data. The Tamura–Nei model (Tamura and Nei 1993) was used to estimate the evolutionary distances between sequences (as recommended for control region sequence data; Kumar et al. 1993) with a gamma correction (a = 0.17) estimated by maximum likelihood using PAUP 4.0b (Swofford 1998). To examine possible differences in gene flow among the sexes, FST estimates were also obtained for males and females separately. Statistical significance of FST and FST estimates was estimated using a permutation simulation with 10,000 permutations of haplotypes among populations and the sequential Bonferroni test was used to correct probability values for multiple comparisons (Rice 1989). The extent of geographic heterogeneity in haplotype frequency distributions was further assessed with Monte Carlo simulations of v2 values as described in Roff and Bentzen (1989) with 10,000 randomizations of the original data matrix using the program REAP (McElroy et al. 1992). This process of performing repeated randomizations of the original data matrix is intended to reduce the effect of large numbers of small cell counts on the validity of the v2 procedure (Roff and Bentzen 1989). Unlike Wright’s F-statistics, which are parameters, v2 is a true test of statistical significance and thus the use of this statistic has been purported to be more powerful at detecting haplotype frequency differences than sequence based statistics (Hudson et al. 1992). To illustrate the relationship among haplotypes, a minimum spanning network was constructed using the program Network 2.0 (Bandelt et al. 1999). Microsatellite genotyping All samples were genotyped at nine dinucleotide microsatellite loci: Ttr04, Ttr11, Ttr19, Ttr34, Ttr48, Ttr63 (Rosel et al. 2005); MK8 (Krützen et al. 2001); EV14, EV37 (Valsecchi and Amos 1996). Amplification reactions contained 25– 50 ng DNA, 20 mM Tris–HCl, pH 8.0, 50 mM KC1, 1.5 mM MgC12, 0.3 lM of each primer, 150 lM dNTPs and 0.75 units of Taq DNA polymerase. The thermal cycler profile consisted of initial denaturation at 94 °C for 30 s, followed by 30 cycles of 94 °C for 20 s, 55–62 °C for 20 s and 72 °C for 40 s, with a final extension at 72 °C for 10 min. All reactions included both positive and negative controls. Following amplification, samples were mixed with an internal size standard and run on either an ABI 310 or ABI 3100 Genetic Analyzer. Genescan Analysis 3.1 and GeneMapper 3.5 software (Applied Biosystems, Foster City, CA) were used for sizing of allele fragments. 719 Microsatellite analyses Allele frequencies, observed heterozygosity (Ho), expected heterozygosity (He), Hardy–Weinberg exact tests (Guo and Thompson 1992) and tests for heterozygote deficiency were carried out using GENEPOP 3.3 (Raymond and Rousset 1995). Estimates of allelic richness per population, a measure of the number of alleles per locus independent of sample size (Petit et al. 1998), were calculated using the program FSTAT (Goudet 2002). To test whether the microsatellite loci were independently inherited, tests for linkage disequilibrium were carried out using Arlequin 2.0 (Schneider et al. 2000). For tests of Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium, the sequential Bonferroni correction was applied to correct probability values for multiple comparisons (Rice 1989). For each area where samples were collected via biopsy darting, individual genotypes were compared against each other to identify possible duplicate samples. The probability of identity calculation from Paetkau et al. (1995) was then used to obtain the probability that two different individuals in our dataset could share the same genotype across all loci. The inbreeding coefficient, FIS (Weir and Cockerham 1984), was estimated for each population using FSTAT where significance values were obtained by permuting alleles 1000 times within populations (Goudet 2002). As a further means of testing whether samples collected from each of the populations contained a significantly high number of related individuals, Queller and Goodnight’s index of relatedness was used to obtain average and pairwise relatedness values for individuals of each population (R; Queller and Goodnight 1989). The distribution of pairwise R values per population was then compared to a distribution generated from the random simulation of 1000 unrelated pairs using each population’s allele frequencies. The simulations were carried out using KINSHIP 1.2 (Goodnight and Queller 1999) and the distributions were statistically compared using a Komogorov–Smirnov two-sample test. The common estimator of population subdivision, FST (Wright 1965), which assumes the infinite allele model (IAM, Crow and Kimura 1970), and a statistic more specific to microsatellite data, RST (Slatkin 1995), which assumes a stepwise mutation model (SMM, Kimura and Ohta 1978), were used to estimate population subdivision using the microsatellite data. Based on the differences in these models, FST takes into account allele frequencies, whereas RST takes into account both allele frequencies and genetic distance. To obtain an FST estimate, an AMOVA was run using the program Arlequin 2.0 (Schneider et al. 2000). As with the mtDNA data, FST estimates were also obtained for males and females separately to examine possible differences in gene flow among the sexes. An unbiased version of RST, RhoST, in which allele sizes are transformed to standard variances, was calculated using the program RST CALC (Goodman 1997). For probability values of both FST and RhoST estimates, the sequential Bonferroni test was used to correct for multiple comparisons (Rice 1989). We used the program MIGRATE 2.0.3 to estimate asymmetric rates of migration between the populations using the microsatellite data set to the Brownian mutation model (Beerli and Felsenstein 2001). MIGRATE uses a coalescentbased maximum likelihood approach to estimate migration assuming constant effective population size while allowing for unequal subpopulation sizes (Beerli and Felsenstein 2001). An initial run used FST based estimates of Nm and h and 10 short chains with 10,000 sampled genealogies and three long chains with 100,000 sampled genealogies (with a burn-in of 10,000). Resulting Nm and h values were then inputted as initial parameters for a second run using 20 short chains with 20,000 sampled genealogies and five long chains with 200,000 sampled genealogies. Results Mitochondrial DNA After all consensus sequences were aligned and tRNA sequence trimmed, a final 359 bp fragment of the mtDNA control region was used for analyses. A total of 15 polymorphic sites defined 11 unique haplotypes among the 223 individuals. One site contained an insertion/deletion while the remaining 14 polymorphic sites were transition mutations. The most common haplotype was shared among the Sarasota Bay, Tampa Bay, Charlotte Harbor and coastal Gulf populations, but this haplotype was not found in the Matag- 720 Table 1. Distribution of mtDNA haplotypes among sampling locations Mitochondrial DNA haplotype Population 1 2 3 4 5 6 7 8 9 10 11 Gulf of Mexico Sarasota Bay Tampa Bay Charlotte Harbor Matagorda Bay 1 0 0 0 0 3 0 1 0 0 18 18 34 30 0 5 0 0 0 0 15 0 0 0 0 8 3 6 7 1 6 15 3 13 10 0 0 1 1 0 0 0 1 0 0 0 0 0 0 16 0 0 0 0 7 Table 2. Total number of females (F) and males (M) sampled per population as well as the total individuals sampled (N) MtDNA genetic diversity estimates +/) standard error. Location F M N Haplotype diversity Nucleotide diversity Gulf of Mexico Sarasota Bay Tampa Bay Charlotte Harbor Matagorda Bay 24 20 17 9 15 32 16 29 42 19 56 36 46 51 34 0.796±0.028 0.586±0.042 0.441±0.085 0.581±0.055 0.668±0.046 0.009±0.005 0.011±0.006 0.007±0.004 0.010±0.006 0.003±0.002 orda Bay population (Table 1). Unique haplotypes were found in the Tampa Bay, Matagorda Bay, and coastal Gulf populations, but not in Sarasota Bay or Charlotte Harbor. Haplotype diversity estimates ranged from 0.441 in Tampa Bay to 0.796 in the coastal Gulf of Mexico. Nucleotide diversity estimates ranged from 0.003 in Matagorda Bay to 0.011 in Sarasota Bay (Table 2). Relationships among the eleven unique haplotypes and their relative frequency are shown as a minimum spanning network in Figure 2. Sequences were deposited into GenBank under Accession numbers AY997307 to AY997311. Results from the AMOVA analyses using the mtDNA sequence data indicated a significant amount of genetic variation among all areas using both frequency information alone (FST = 0.186, P < 0.001), and haplotype frequency and genetic distance information combined (FST = 0.323, P < 0.001). After sequential Bonferroni correction, FST values showed significant differentiation for all pairwise population comparisons except for the comparisons of Sarasota Bay to Charlotte Harbor and Tampa Bay to Charlotte Harbor (Table 3). Although fewer pairwise comparisons were found to be significant when both haplotype frequency and genetic distance data were combined (FST), the overall partitioning of the genetic variation among areas was greater using this estimate. Based on v2 randomizations, the observed Figure 2. Minimum spanning network showing the relationships among 11 bottlenose dolphin haplotypes. Haplotype numbers correspond to those listed in Table 1. The size of each circle approximates the relative frequency of the haplotype in the total sample. Each hash mark represents one mutational event. 721 Table 3. Estimates of population differentiation for all pairwise population comparisons for mtDNA and microsatellite data where *P<0.05, **P<0.01, ***P<0.001 following sequential Bonferroni correction Populations SB–TB SB–CHB TB–CHB GOM–SB GOM–TB GOM–CHB MTB–SB MTB–TB MTB–CHB MTB–GOM Mitochondrial DNA Microsatellites FST FST v2 FST RST 0.137*** 0.008 0.036 0.113*** 0.155*** 0.098** 0.284*** 0.441*** 0.325*** 0.236*** 0.185*** 0.005 0.073 0.115** 0.062 0.051 0.491*** 0.753*** 0.553*** 0.642*** 15.94* 3.25 8.34 27.06* 29.51* 30.48* 42.98* 67.05* 57.80* 70.27* 0.027*** 0.032*** 0.024*** 0.042*** 0.030*** 0.026*** 0.043*** 0.042*** 0.032*** 0.050*** 0.003 0.008 0.022** 0.030** 0.056*** 0.022** 0.048** 0.060*** 0.035** 0.029** SB: Sarasota Bay, TB: Tampa Bay, CHB: Charlotte Harbor, GOM: coastal Gulf of Mexico, MTB: Matagorda Bay. distribution of haplotypes was significantly heterogeneous for all pairwise population comparisons except for the comparisons of Sarasota Bay to Charlotte Harbor and Tampa Bay to Charlotte Harbor (Table 3). Sample sizes per sex, as determined either in the field or using genetics, are shown in Table 2. When analyzed separately, both males and females showed significant differentiation (FST = 0.153, P < 0.001 and FST = 0.240, P < 0.001, respectively). Microsatellite results The number of alleles per microsatellite locus ranged from 2 for Ttr19 in the Sarasota Bay population to 20 for EV37 in the coastal Gulf population. No locus showed significant evidence of genotypic disequilibrium nor deviated significantly from Hardy–Weinberg equilibrium (HWE) in any population following sequential Bonferroni correction. The average number of alleles per locus was highest in the coastal Gulf population and lowest in Sarasota Bay (Table 4). Out of a total of 27 unique alleles, eleven were found in the coastal Gulf population, nine were found in Matagorda Bay, five were found in Charlotte Harbor, and two were found in Tampa Bay. For eight out of nine loci, all populations shared the most common allele except the Matagorda Bay population which had a different most common allele at four of the nine loci. Average observed heterozygosity values ranged from 0.580 in Matagorda Bay to 0.700 in the coastal Gulf of Mexico (Table 4). FIS estimates and their corresponding P-values showed no indication of significant inbreeding in any population (Table 5). From the probability of identity estimates, the probability of two dolphins sharing the same genotype across all nine loci was highest for Sarasota Bay (3.710)8) and lowest for the coastal Gulf of Mexico (6.210)10, Table 4). In total, nine pairs of biopsy samples were identified as having identical genotypes across all nine loci. Based on the low probability of identity values for all populations, one of each of these pairs was considered to be a duplicate sample and removed from all analyses. Average relatedness values per population ranged from )0.005 to 0.001 (Table 5). Results from the comparison of pairwise R value distributions per population to distributions generated from the Table 4. Estimates of the average number of alleles (A), allelic richness (RS), observed heterozygosity (Ho), expected heterozygosity (He), and probability of identity (PID) per population Gulf of Mexico Sarasota Bay Tampa Bay Charlotte Harbor Matagorda Bay A Rs Ho He P(ID) 9.4 6.4 8.1 8.4 7.3 8.6 6.4 7.6 7.7 7.3 0.700 0.627 0.597 0.601 0.580 0.720 0.636 0.627 0.633 0.624 6.210)10 3.710)8 3.510)8 1.210)8 1.710)8 722 Table 5. FIS values and corresponding significance values estimated by FSTAT, average relatedness estimates (R) per population +/) the standard error and significance values from Kolmogorov–Smirnov two-sample tests comparing the distribution of pairwise R values per population to the distribution generated from a random simulation of 1000 pairs using the population allele frequencies Location FIS P-value R ± SE P-value Gulf of Mexico Sarasota Bay Tampa Bay Charlotte Harbor Matagorda Bay 0.022 0.005 0.038 0.039 0.053 0.17 0.48 0.09 0.06 0.06 )0.001 ± 0.005 )0.005 ± 0.009 0.001 ± 0.007 )0.002 ± 0.005 )0.004 ± 0.009 0.06 0.36 0.37 0.24 0.76 Table 6. MIGRATE asymmetric migration rates (4Nem) Source population GOM SB TB CHB MTB Receiving population GOM SB TB CHB MTB – 16.7 7.5 11.2 14.1 25.2 – 22.4 17.5 8.5 9.7 23.3 – 17 11.4 22.7 12 11.3 – 14.7 20.8 7.7 10.1 13.7 – SB: Sarasota Bay, TB: Tampa Bay, CHB: Charlotte Harbor, GOM: coastal Gulf of Mexico, MTB: Matagorda Bay. random simulation of 1000 unrelated pairs showed that no population contained a significantly higher number of related individuals than expected in a randomly mating population (Table 5). AMOVA results from the microsatellite data indicate significant genetic structure among all five populations (FST = 0.034, P < 0.0001; RhoST = 0.032, P < 0.0001) and all pairwise population comparisons resulted in significant FST values (Table 3). As with the mtDNA data, when analyzed separately, both males and females showed significant differentiation (FST = 0.032, P < 0.001 and FST = 0.046, P < 0.001, respectively). All pairwise population comparisons were also significant using the RhoST estimator except for the comparisons of Sarasota Bay to Tampa Bay and Sarasota Bay to Charlotte Harbor (Table 3). Asymmetrical migration rate estimates produced by MIGRATE ranged from 7.5 to 25.2 (4Nem, Table 6). Discussion Population structure Results from this study reveal fine scale geographic structure of bottlenose dolphins in the Gulf of Mexico. Based on the mtDNA data, when haplo- type frequency and genetic distance are considered, approximately 32% of the total variation detected can be attributed to differences among these populations (FST = 0.323). In context of other mtDNA studies, this level of differentiation is similar to that found among beluga whales of five summer areas in Alaska and north–west Canada (O’ Corry-Crowe et al. 1997), and greater than that found among many other cetacean populations (Dizon et al. 1994; Rosel et al. 1999a; Escorza-Trevino and Dizon 2000). When just haplotype frequency information was considered for population structure analyses, the percentage of variation attributed to differences among populations was reduced but remained significant (FST 0.186, P < 0.001). The FST estimator is strongly influenced by the presence of the geographically isolated Matagorda Bay samples. These samples are dominated by a haplotype not found in any of the eastern populations and this haplotype differs from the most common haplotype in the other populations by a relatively high number of nucleotide substitutions (see Figure 2). The microsatellite data also revealed significant differentiation between Matagorda Bay and all Florida inshore populations. These results suggest there has been very little recent genetic exchange between the Matagorda Bay and eastern Gulf populations. Interestingly, from the minimum spanning net- 723 work, it appears that two of the three most common haplotypes (6 and 7) among the Florida inshore populations are genetically distant from the third (3), and appear to be more closely related to the two haplotypes unique to Matagorda Bay (10 and 11) than the other haplotypes found in the eastern Gulf. One explanation would be that haplotypes 3, 6, and 7 were the most common haplotypes of the source population from which these inshore populations were founded (discussed below). The fact that the less common haplotypes in the coastal Gulf, Tampa Bay and Charlotte Harbor populations appear to have radiated from the most common haplotype in these populations, haplotype 3, also supports this hypothesis. Excluding Matagorda Bay, significant population differentiation was also found solely among the eastern Gulf sampling sites. However, the mtDNA and microsatellite data differ in their estimates for two inshore population pairwise comparisons: Sarasota Bay to Charlotte Harbor and Tampa Bay to Charlotte Harbor (Table 3). Although one possible explanation for this could be female-biased dispersal, higher overall FST estimates for females alone (vs. males) and differentiation detected with both mtDNA and nuclear markers for all other pairwise comparisons argue against this. Instead, two characteristics of the data may explain the discrepancies between the mtDNA and nuclear population structure estimates. First, low haplotype variability (11 unique haplotypes in 223 samples) compared to generally high microsatellite diversity may translate to the microsatellites being a more sensitive indicator of population structure. The low level of variability in the mtDNA sequences could be an artifact of recent colonization of the inshore bays following a founder event (discussed below). This effect of low haplotypic diversity on the population structure estimates helps to explain why previous studies of population structure of bottlenose dolphins in this area, using RFLP analysis of mtDNA markers (Duffield and Wells 1991, 2002; Dowling and Brown 1993), did not detect the differences reported here using microsatellite loci. A second factor that is likely contributing to these results is the bias in sample size for males (n = 42) vs. females (n = 9) in the Charlotte Harbor sample. Not only is much power lost in the very small sample size for females but, concurrently, if some males are moving between populations, the effect of sampling a male from a neighboring area and erroneously including it in the Charlotte Harbor sample would further reduce the accuracy of genetic diversity estimates in the population, thereby effecting estimates of genetic structure. In the absence of additional female samples from this population, we cannot rule out the effect this bias is having on the population comparisons involving Charlotte Harbor. These discrepancies between mtDNA and microsatellite FST estimates illustrate the importance of utilizing more than one type of marker when making inferences about the genetic structure of populations. The significant differentiation detected between all eastern Gulf sampling sites using the microsatellite data is surprising given the relative mobility of these dolphins, the close proximity of these populations, the evidence for mixed groups (Wells 1986; Fazioli 1999) and reported movements of animals (both males and females) between these different areas (Duffield and Wells 1991). Wells (1986) and Fazioli (1999) have documented mixed groups containing both inshore and coastal bottlenose dolphins in these areas. Though Wells (1986) did note that, for Sarasota Bay dolphins, mixing was more frequent outside of the primary breeding season. While mixed groups may provide significant opportunity for genetic exchange to occur between the coastal Gulf and Sarasota Bay populations, genetic results indicate that, although the opportunity may be present, little interbreeding is occurring. Similar results have been found for other cetacean species where significant population subdivision was detected despite observational data documenting a high occurrence of mixed groups (in beluga whales, Brennin et al. 1997; in humpback whales, Valsecchi et al. 1997 and in Tursiops sp., Wang et al. 1999). One potential concern is that the population structure detected here among the inshore populations reflects a high degree of relatedness among sampled animals within each bay rather than true population structure between bays. Genealogy data are only available for Sarasota Bay dolphins and, based on these data, we can say that no known first-order relatives were included in that dataset. Tests for significant FIS values did not reveal evidence for inbreeding in any of the populations sampled. Furthermore, it is unlikely that these bays simply contain a single family group of dolphins that has moved into the area for only a 724 generation or two. The observation of at least four generations in Sarasota Bay is documented (Duffield and Wells 2002) and the abundance estimate of 120 Sarasota Bay resident dolphins given by Wells (2003) is too large to be comprised of a single family group. Finally, in examining the distributions of relatedness values for each population, we found no evidence for significantly elevated R values within populations, as no population had a distribution that differed significantly from that of a randomly generated population of unrelated individuals (Table 5). Taken together, this evidence indicates that these are populations subjected to evolutionary forces rather than family groups occupying the inshore habitats for only a generation or two. Population history Overall, the coastal Gulf population had the highest genetic diversity as measured by haplotypic diversity, allelic richness, and microsatellite heterozygosities. Although abundance estimates are not available for all areas, it is generally thought that the coastal Gulf population is larger in size than the resident inshore populations. Wells (2003) estimates 120 resident dolphins currently inhabit Sarasota Bay, while abundance for bottlenose dolphins in the eastern coastal Gulf of Mexico stock (between 84° W longitude and Key West, Florida) has been estimated at 9912 dolphins (Waring et al. 2002). The larger population in the coastal Gulf would be able to maintain a higher genetic diversity than the smaller, insular inshore populations. It is also possible that the inshore bottlenose dolphin populations along the central west coast of Florida represent founder populations originating from the larger coastal population. This hypothesis would help to explain the fact that the majority of unique haplotypes and microsatellite alleles were found in the coastal Gulf rather than the inshore populations. That is, the inshore populations would carry only a small proportion of the genetic variation from the larger source population, as certain alleles and haplotypes that were present in the larger population would have been lost during the founding events. The founder hypothesis has been suggested by other studies investigating population structure of bottlenose dolphins in the Gulf of Mexico (Natoli et al. 2004) and western North Atlantic (Hoelzel et al. 1998). The fact that no fixed differences were found between the eastern inshore and coastal Gulf populations suggests that either isolation was a relatively recent event and/or a low level of gene flow is still occurring among them. From geological data, it is clear that the formation of the inshore bays and estuaries along Florida’s west coast, and therefore the founding of these areas by bottlenose dolphins, were relatively recent events (Davis 1997). This may explain the lack of fixed differences between the eastern inshore and coastal Gulf populations. Alternatively, low levels of genetic exchange may also be ongoing. Indirect estimates of migration from the genetic data are concordant with those documented from observational data (Wells 1986) and relatively low, but non-zero. Migration estimates presented here do need to be taken with caution as certain assumptions of the underlying models, such as that of constant effective population size and sampling of all source populations, are potentially violated with the data at hand. Nevertheless, private alleles are seen in four of the five populations suggesting that the low level of gene flow is not enough to homogenize these populations. Resource specialization and social structure While founder events may have originally established the inshore populations, what has maintained their isolation from the coastal Gulf? We suggest differences in habitat and resource use are mechanisms by which this differentiation has evolved and/or is maintained. Although the inshore dolphins have easy access to the nearby more open coastal Gulf waters, they spend the majority of their time in the shallower, more confined waters of the bays and estuaries. It may be that these areas provide more protection from predators (sharks) or that particular prey species are more abundant in the inshore habitats (Wells et al. 1980; Connor et al. 2000). A certain degree of resource specialization with respect to prey has been documented for the Sarasota Bay and coastal Gulf populations. In a Barros and Wells (1998) study, the distribution of Sarasota Bay feeding events was correlated with prey habitat distribution and results indicated that feeding behaviors were most often observed near seagrasses. As 725 pinfish are often the dominant fish species in Florida seagrass meadows (Stoner 1983), and a primary prey species of Sarasota resident dolphins (Barros and Wells 1998), one likely factor contributing to the distribution of resident Sarasota Bay dolphins is the distribution of the seagrass meadows harboring their primary prey. In contrast, analysis of stomach contents of dolphins stranded on Gulf coastal beaches indicates that, from 1985 to 1996, 21% of the dolphins consumed cephalopods, a prey item almost entirely absent in the Sarasota Bay residents (Barros and Wells 1998). Similar differences have been found for bottlenose dolphin populations on the east coast of Florida where dolphins in the Indian River Lagoon system are almost exclusively piscivorous, whereas dolphins found in the adjacent coastal waters have a mixed diet of fish and squid (Barros 1993). These differences, though not as extreme, are not unlike differences found in the diets of resident and transient killer whales (Matkin and Saulitis 1994). Hoelzel and Dover (1991) suggested that differing foraging strategies between sympatric killer whale populations served as a behavioral isolating mechanism by which genetic differences evolved. As with killer whales, resource specialization with respect to prey is most likely a key factor contributing to the separation of inshore and coastal bottlenose dolphin populations. Among the inshore areas, it is likely that the complex social structure of these dolphins is contributing to heterogeneity. Long-term social relationships documented in bottlenose dolphins include the formation of male pairs and female bands in Sarasota Bay (Wells 1991; Owen et al. 2002; Wells 2003) and first-order and second-order male alliances in Shark Bay, Western Australia (Connor et al. 1992). Still some bottlenose dolphins in these populations, both males and females, appear to remain more solitary, without forming any strong social relationships (Wells 1991). Given these social affiliations, it is not surprising that our genetic data support previous data from field observations, that bottlenose dolphins in these populations exhibit site philopatry (Wells 1991). Though male-biased dispersal is well documented in many other cetacean species (Baker et al. 1993; Rosel et al. 1999a; Escorza-Trevino and Dizon 2000), and recently in bottlenose dolphins (T. aduncus Möller et al. 2004), our results indicate that both males and females exhibit some degree of natal site philopatry, as indicated by significant FST values for each sex based on both mtDNA and nuclear microsatellite data. These results are similar to those reported for bottlenose dolphins in Shark Bay, Western Australia, where males have been shown to disperse farther than females yet still exhibit natal philopatry (Tursiops sp. Krützen et al. 2004) and, more recently, to a global study of bottlenose dolphins where significant structure was detected using both mtDNA and nuclear markers (Natoli et al. 2004). Management implications Current threats to bottlenose dolphins in the Gulf of Mexico include entanglement in recreational and commercial fishing gear, large scale die-offs due to epizootic events and undetermined causes, environmental contaminants and other pollutants, and disturbance and collisions from increasing boat traffic in coastal areas. The anthropogenic threats mentioned are especially relevant to the inshore dolphin populations on the central west coast of Florida where human population size, development, and boat traffic are increasing at an alarming rate. Although bottlenose dolphins are not endangered, the long-term research conducted on these populations allows a unique opportunity to monitor not only the health of these populations, but also the health of their habitat and ecosystem over long periods of time. The genetic data presented here indicate that inshore waters do harbor semi-isolated populations of bottlenose dolphins. As a result, ecosystem health and monitoring of a single location used as a proxy for other adjacent areas may not be adequate. Furthermore, the relatively low estimates of migration presented here, coupled with a generation time of approximately 10 years when estimated as age at first reproduction (Wells 2003), means that, should one of these populations be lost, recolonization to genetically healthy population sizes could take on the order of several hundred years. Such an event would violate the mandate under the US Marine Mammal Protection Act requiring that populations remain functioning elements of their ecosystems. Thus, genetic results from this study support the current identification of separate management units (Waring et al. 2002) for different embayments in inshore areas of the Gulf of 726 Mexico. These results also indicate the occurrence of a separate eastern coastal Gulf stock for dolphins inhabiting the nearshore waters along the west coast of Florida. Whether the current putative eastern coastal Gulf stock is comprised of more than one population needs to be examined. In addition, future studies will reveal whether similar genetic differentiation exists among the other 29 defined stocks of bottlenose dolphins in the bays, sounds and estuaries of the Gulf of Mexico, and whether some of the larger bays may harbor more than one genetically distinct stock or population. Acknowledgments We are grateful to the following people for their assistance with sample collection: E. Zolman, J. Thera, M. Alsina, T. Speakman, K. Fazioli, A. Westgate, K. Hull, S. Hofmann and the entire staff of the Sarasota Dolphin Research Program. This work also benefited from discussions with and assistance from L. Adams, S. Kingston, K. Tolley, N. Barros, M. Krützen, L. Schwacke, A. Strand and D. Queller. Funding and/or logistical support for this research was provided by the Florida Fish and Wildlife Conservation Commission, Chicago Zoological Society, Dolphin Quest, Mote Marine Laboratory, Sarasota Dolphin Research Program and the U. S. National Marine Fisheries Service (NMFS). Biopsy sampling was carried out under NMFS Scientific Research Permits No. 522-1527 and No. 522-1569. References Baker CS, Perry A, Bannister JL, Weinrich MT, Abernathy RB (1993) Abundant mitochondrial DNA variation and worldwide population structure in humpback whales. Proc. Natl Acad. Sci. USA, 90, 8239–8243. Bandelt HJ, Forster P, Röhl A (1999) Median joining networks for inferring intraspecific phylogenies. Mol. Biol. Evol., 16, 37–48. Barros NB (1993) Feeding Ecology and Foraging Strategies of Bottlenose Dolphins on the Central East Coast of Florida. PhD thesis, University of Miami. Barros NB, Odell DK (1990) Food habits of bottlenose dolphins in the southeastern United States. In: The Bottlenose Dolphin (eds. Leatherwood S, Reeves RR), pp. 309–328. Academic Press, San Diego. Barros NB, Wells RS (1998) Prey and feeding patterns of resident bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, FL. J. Mammal., 79, 1045–1059. Beerli P, Felsenstein J (2001) Maximum likelihood estimation of a migration matrix and effective population sizes in n subpopulations by using a coalescent approach. Proc. Natl Acad. Sci. USA, 98, 4563–4568. Brennin R, Murray BW, Friesen MK, et al. (1997) Population genetic structure of beluga whales (Delphinapterus leucas): Mitochondrial DNA sequence variation within and among North American populations. Can. J. Zool., 75, 795–802. Connor RC, Smolker RA, Richards AF (1992) Two levels of alliance formation among male bottlenose dolphins (Tursiops sp.). Proc. Natl Acad. Sci. USA, 89, 987–990. Connor RC, Wells RS, Mann J, Read AJ (2000) The bottlenose dolphin: Social relationships in a fission–fusion society. In: Cetacean Societies: Field Studies of Dolphins and Whales (eds. Mann J, Connor RC, Tyack PL, Whitehead H), pp. 91– 126. University of Chicago Press, Chicago. Crow JF, Kimura M (1970) An Introduction to Population Genetics Theory, Harper and Row New York, Evanston and London. Davis RA (1997) Geology of the Florida coast In: The Geology of Florida (eds. Randazzo AF, Jones DS), pp. 155–168. The University Press of Florida, Gainesville. Dizon AE, LeDuc CA, LeDuc RG (1994) Intraspecific structure of the northern right whale dolphin (Lissodelphis borealis): The power of an analysis of molecular variation for differentiating genetic stocks. Calif. Coop. Ocean. Fish. Invest. Rep., 35, 61–67. Dowling TE, Brown WM (1993) Population structure of the bottlenose dolphin (Tursiops truncatus) as determined by restriction fragment length polymorphism analysis of mitochondrial DNA. Mar. Mamm. Sci., 9, 138–155. Duffield DA, Wells RS (1991) The combined application of chromosome, protein and molecular data for the investigation of social unit structure and dynamics in Tursiops truncatus. In: Genetic Ecology of Whales and Dolphins (eds. Hoelzel AR, Donovan GP), pp. 155–169. International Whaling Commission, Cambridge. Duffield DA, Wells RS (2002) The molecular profile of a resident community of bottlenose dolphins, Tursiops truncatus. In: Molecular and Cell Biology of Marine Mammals (eds. Pfeiffer CJ), pp. 3–11. Kreiger Publishing Company, Malabar. Escorza-Trevino S, Dizon AE (2000) Phylogeography, intraspecific structure and sex-biased dispersal of Dall’s porpoise, Phocoenoides dalli, revealed by mitochondrial and microsatellite analyses. Mol. Ecol., 9, 1049–1060. Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics, 131, 479–491. Fazioli KL (1999) Distribution, Relative Abundance, and Community Structure of Coastal Bottlenose Dolphins (Tursiops truncatus) in the Gulf of Mexico off Sarasota, FL. Master of Science thesis, University of California, Santa Cruz. 727 Gilbert DG (1995) Seqpup, Biosequence Editor and Analysis Platform, Version 0.6. Bionet Software Available at: http:// www.hgmp.mrc.ac.uk/Registered/Option/segpup.html. Goodman SJ (1997) RSTCALC: A collection of computer programs for calculating estimates of genetic differentiation from microsatellite data and determining there significance. Mol. Ecol., 6, 881–885. Goodnight KF, Queller DC (1999) Computer software for performing likelihood tests of pedigree relationship using genetic markers. Mol. Ecol., 8, 1231–1234. Goudet J (2002) FSTAT, a Program to Estimate and Test Gene Diversities and Fixation Indices Version 2.9.3. Available from http://www.unil.ch/izea/softwares/fstat.html. Guo SW, Thompson EA (1992) Performing the exact test of Hardy–Weinberg proportion for multiple alleles. Biometrics, 48, 361–372. Hersh SL, Duffield DA (1990) Distinction between Northwest and Atlantic offshore and coastal bottlenose dolphins based on hemoglobin profile and morphometry. In: The Bottlenose Dolphin (eds. Leatherwood S, Reeves RR), pp. 129–139. Academic Press, San Diego. Hoelzel AR (1998) Genetic structure of cetacean populations in sympatry, parapatry, and mixed assemblages: Implications for conservation policy. J. Hered., 89, 451–458. Hoelzel AR, Dover GA (1991) Genetic differentiation between sympatric killer whale populations. Heredity, 66, 191–195. Hoelzel AR, Potter CW, Best PB (1998) Genetic differentiation between parapatric ‘nearshore’ and ‘offshore’ populations of the bottlenose dolphin. P. Roy. Soc. Lon. B Bio., 265, 1177– 1183. Hudson RR, Boos DD, Kaplan NL (1992) A statistical test for detecting geographic subdivision. Mol. Biol. Evol., 9, 138–151. Irvine B, Wells RS (1972) Results of attempts to tag Atlantic bottlenose dolphins (Tursiops truncatus). Cetology, 13, 1–5. Kingston SE, Rosel PE (2003) Genetic differentiation among recently diverged delphinid taxa determined using AFLP markers. J. Hered., 95, 1–10. Kimura M, Ohta T (1978) Stepwise mutation model and distribution of allelic frequencies in a finite population. Proc. Natl Acad. Sci. USA, 75, 2868–2872. Krützen M, Valsecchi E, Connor R, Sherwin WB (2001) Characterization of microsatellite loci in Tursiops aduncus. Mol. Ecol. Notes, 1, 170–172. Krützen M, Sherwin WB, Berggren P, Gales N (2004) Population structure in an inshore cetacean revealed by microsatellite and mtDNA analysis: Bottlenose dolphins (Tursiops sp.) in Shark Bay, Western Australia. Mar. Mamm. Sci., 20, 28–47. Kumar S, Tamura K, Nei M (1993) MEGA: Molecular Evolutionary Genetics Analysis, Version 10, The Pennsylvania State University, University Park, PA16802. Matkin CO, Saulitis EL (1994) Killer Whale (Orcinus orca) Biology and Management in Alaska, Marine Mammal Commission Report T75135023. McElroy D, Moran P, Bermingham E, Kornfield I (1992) REAP: The restriction enzyme analysis package. J. Hered., 83, 157–158. Mead JG, Potter CW (1995) Recognizing two populations of the bottlenose dolphin (Tursiops truncatus) off the Atlantic coast of North America: Morphologic and ecologic considerations. IBI Rep., 5, 31–44. Möller LM, Beheregaray L (2004) Genetic evidence for sexbiased dispersal in resident bottlenose dolphins (Tursiops truncatus). Mol. Ecol., 13, 1607–1612. Natoli A, Peddemors VM, Hoelzel AR (2004) Population structure and speciation in the genus Tursiops based on microsatellite and mitochondrial DNA analyses. J. Evol. Biol., 17, 363–375. Nei M (1987) Molecular Evolutionary Genetics, Columbia University Press, New York. O’Corry-Crowe GM, Suydam RS, Rosenberg A, Frost KJ, Dizon AE (1997) Phylogeography, population structure and dispersal patterns of the beluga whale Delphinapterus leucas in the western Neararctic revealed by mitochondrial DNA. Mol. Ecol., 6, 955–970. Owen ECG, Hofmann S, Wells RS (2002) Ranging and social association patterns of paired and unpaired adult male bottlenose dolphins, Tursiops truncatus, in Sarasota, Florida, provide no evidence for alternative mating strategies. Can. J. Zool., 80, 2072–2089. Paetkau D, Calvert W, Stirling I, Strobeck C (1995) Microsatellite analysis of population structure in Canadian polar bears. Mol. Ecol., 4, 347–154. Petit RJ, El Mousadik A, Pons O (1998) Identifying populations for conservation on the basis of genetic markers. Conserv. Biol., 12, 844–855. Queller DC, Goodnight KL (1989) Estimating relatedness using genetic markers. Evolution, 43, 258–275. Raymond M, Rousset F (1995) GENEPOP (Version 1.2). Population genetics software for exact tests and ecumenism. J. Hered., 86, 248–249. Rice WR (1989) Analyzing tables of statistical tests. Evolution, 43, 223–225. Roff DA, Bentzen P (1989) The statistical analysis of mitochondrial DNA polymorphisms: v2 and the problem of small samples. Mol. Biol. Evol., 6, 539–545. Rosel PE (2003) PCR-based sex determination in Odontocete cetaceans. Conserv. Genet., 4, 647–649. Rosel PE, Block BA (1996) Mitochondrial control region variability and global population structure in the swordfish, Xiphias gladius. Mar. Biol., 125, 11–22. Rosel PE, Forgetta V, Dewar K (2005) Isolation and characterization of twelve polymorphic microsatellite markers in bottlenose dolphins (Tursiops trunctatus). Mol. Ecol. Notes, (doi: 10.1111/j.1471-8286-2005.01078.x) in press. Rosel PE, France SC, Wang JY, Kocher TD (1999a) Genetic structure of harbour porpoise Phocoena phocoena populations in the northwest Atlantic based on mitochondrial and nuclear data. Mol. Ecol., 8, S41–S54. Rosel PE, Tiedemann R, Walton M (1999b) Genetic evidence for limited trans-Atlantic movements of the harbor porpoise, Phocoena phocoena. Mar. Biol., 133, 583–591. Ross GJB, Cockcroft VC (1990) Comments on Australian bottlenose dolphins and the taxonomic status of Tursiops aduncus. In: The Bottlenose Dolphin (eds. Leatherwood S, Reeves RR), pp. 101–128. Academic Press, San Diego. Schneider S, Roessli D, Excoffier L (2000) Arlequin: A Software Package for Population Genetic Data Analysis, Version 2.0, 728 Genetics and Biometry Laboratory, University of Geneva, Switzerland. Scott MD, Wells RS, Irvine AB (1990) A long-term study of bottlenose dolphins on the west coast of Florida. In: The Bottlenose Dolphin (eds. Leatherwood S, Reeves RR), pp. 235–244. Academic Press, San Diego. Shane SH, Wells RS, Würsig B (1986) Ecology, behavior, and social organization of the bottlenose dolphin: A review. Mar. Mamm. Sci., 2, 34–63. Skúlason S, Smith TB (1995) Resource polymorphism in vertebrates. TREE, 10, 366–370. Slatkin M (1995) A measure of population subdivision based on microsatellite allele frequencies. Genetics, 139, 457–462. Stoner AW (1983) Distribution of fishes in seagrass meadows: Role of macrophyte biomass and species composition. Fish. Bull., 81, 837–847. Swofford DL (1998) PAUP*. Phylogenetic Analysis using Parsimony (*and Other Methods). Version 4.0b, Sinauer Associates, Sunderland, Massachusetts. Tamura K, Nei M (1993) Estimation of the number of nucleotide substitutions in the control region of mtDNA in humans and chimpanzees. Mol. Biol. Evol., 10, 512–526. Valsecchi E, Amos W (1996) Microsatellite markers for the study of cetacean populations. Mol. Ecol., 5, 151–156. Valsecchi E, Palsbøll P, Hale P, Glockner-Ferrari D, Ferrari M, Clapham P, Larsen F, Mattila D, Sears R, Sigurjonsson J, Brown M, Corkeron P, Amos B (1997) Microsatellite genetic distances between oceanic populations of the humpback whale (Megaptera novaeangliae). Mol. Biol. Evol., 14, 355–362. Wang JY, Chou LS, White BN (1999) Mitochondrial DNA analysis of sympatric morphotypes of bottlenose dolphins (genus: Tursiops) in Chinese waters. Mol. Ecol., 8, 1603–1612. Waring GT, Quintal JM, Fairfield C (2002) US. Atlantic and Gulf of Mexico Stock Assessments. NOAA Technical Memorandum. NMFS-NE-169. Weigle B (1990) Abundance, distribution, and movements of bottlenose dolphins (Tursiops truncatus) in Lower Tampa Bay, FL In: Individual Recognition of Cetaceans: Use of Photo-Identification and Other Techniques to Estimate Population Parameters (eds. Hammond PS, Mizroch SA, Donovan GP), pp. 195–201. International Whaling Commission, Cambridge. Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population structure. Evolution, 38, 1358–1370. Wells RS (1986) Population Structure of Bottlenose Dolphins: Behavioral Studies Along the Central West Coast of Florida. Contract Report to the National Marine Fisheries Service, Southeast Fisheries Center. Contract No. 45-WCNF-500366. Wells RS (1991) The role of long-term study in understanding the social structure of a bottlenose dolphin community In: Dolphin Societies: Discoveries and Puzzles (eds. Pryor K, Norris KS), pp. 199–225. University of California Press, Berkeley. Wells RS (2003) Dolphin social complexity: Lessons from longterm study and life history In: Animal Social Complexity: Intelligence, Culture, and Individualized Societies (eds. Waal FBMde, Tyack PL), pp. 32–56. Harvard University Press, Cambridge. Wells RS, Irvine AB, Scott MD (1980) The social ecology of inshore odontocetes. In: Cetacean Behavior: Mechanisms and Processes (eds. Herman LM), pp. 263–317. Wiley & Sons, New York. Wells RS, Bassos KM, Urian KW, Carr WJ, Scott MD (1996a) Low-level Monitoring of Bottlenose Dolphins, Tursiops truncates, in Charlotte Harbor, Florida 1990–1994. NOAA Technical Memorandum. NMFS-SEFSC-384. Wells RS, Urian KW, Read AJ, Bassos MK, Carr WJ, Scott MD (1996b) Low-Level Monitoring of Bottlenose Dolphins, Tursiops truncatus, in Tampa Bay, Florida 1988-1993. NOAA Technical Memorandum. NMFS-SEFSC-385. Wright S (1965) The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution, 19, 395–420. Würsig B, Lynn SIC (1996) Movements, Site Fidelity, and Respiration Patterns of Bottlenose Dolphins on the Central Texas Coast. NOAA Technical Memorandum NMFS-SEFSC-383. Zolman ES (2002) Residence patterns of bottlenose dolphins (Tursiops truncatus) in the Stono River estuary, Charleston County, South Carolina, USA. Mar. Mamm. Sci., 18, 879– 892.