The scalloped hammerhead shark, Sphyrna lewini, is a large endangered predator with a circumglobal distribution, observed in the open ocean but linked ontogenetically to coastal embayments for parturition and juvenile development. A previous survey of maternal (mtDNA) markers demonstrated strong genetic partitioning overall (global ΦST = 0.749) and significant population separations across oceans and between discontinuous continental coastlines.
We surveyed the same global range with increased sample coverage (N = 403) and 13 microsatellite loci to assess the male contribution to dispersal and population structure. Biparentally inherited microsatellites reveal low or absent genetic structure across ocean basins and global genetic differentiation (FST = 0.035) over an order of magnitude lower than the corresponding measures for maternal mtDNA lineages (ΦST = 0.749). Nuclear allelic richness and heterozygosity are high throughout the Indo-Pacific, while genetic structure is low. In contrast, allelic diversity is low while population structure is higher for populations at the ends of the range in the West Atlantic and East Pacific.
These data are consistent with the proposed Indo-Pacific center of origin for S. lewini, and indicate that females are philopatric or adhere to coastal habitats while males facilitate gene flow across oceanic expanses. This study includes the largest sampling effort and the most molecular loci ever used to survey the complete range of a large oceanic predator, and findings emphasize the importance of incorporating mixed-marker analysis into stock assessments of threatened and endangered shark species.
Citation: Daly-Engel TS, Seraphin KD, Holland KN, Coffey JP, Nance HA, Toonen RJ, et al. (2012) Global Phylogeography with Mixed-Marker Analysis Reveals Male-Mediated Dispersal in the Endangered Scalloped Hammerhead Shark (Sphyrna lewini). PLoS ONE 7(1): e29986. doi:10.1371/journal.pone.0029986
Editor: Vincent Laudet, Ecole Normale Supérieure de Lyon, France
Received: August 10, 2011; Accepted: December 9, 2011; Published: January 10, 2012
Copyright: © 2012 Daly-Engel et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Funding came from the Ecology, Evolution and Conservation Biology (EECB) Program at the University of Hawaii, the National Science Foundation (NSF Graduate K-12 program grant to EECB No. 0232016, OCE-0453167, OCE-0623678, and OCE-0929031 to BB and RT, and EPS-0554657 to University of Hawaii), the Zoology Department at the University of Hawaii, the American Association of University Women, and by grant No. 2 K12 GM000708 to the PERT Program at the University of Arizona from the National Institutes of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
As with other sexually reproducing species, reproductive behavior in sharks has implications for population structure. Females make the greater investment in reproduction with large ova, long gestation times, and time spent transiting to coastal nursery grounds. Males contribute less energy to reproduction and are expected to exhibit promiscuity , , . The discrepancy between male and female optimal fitness strategies can produce behaviors that influence genetic architecture, including sex-biased dispersal . Previous work in vertebrates has shown mammals to be largely male-biased in dispersal, with females undergoing limited dispersal due to higher site fidelity. This pattern is consistent among highly migratory marine mammals, as recently demonstrated for sperm whales (Physeter macrocephalus) in the North Atlantic . In birds, however, the opposite is largely true: male birds establish and defend territories, whereas females move between them, exhibiting female-biased dispersal . Because shark reproduction more closely resembles that of marine mammals than other fishes , theoretical expectations are that male-biased dispersal may predominate in this group.
Population structure of coastal-oceanic sharks is something of an enigma, conforming neither to the expectations of sedentary coastal fishes with pelagic larvae, nor the oceanic migrants such as billfish and tunas. Unlike most teleost (bony) fishes, sharks are viviparous, producing small numbers of highly developed young that are capable of swimming and navigation soon after parturition. As a result, many shark species are dependent on shallow coastal habitat for birthing and offspring development , , . The dichotomy of long-range dispersal ability and coastal reproductive habitat may result in complex population genetic structure, in which female and male (or biparental) markers demonstrate contrasting geographic partitions , .
Sampling large marine predators is challenging, and few studies have documented phylogeography on a global scale, most of these focusing exclusively on marine mammals (but see , , , ). Only three studies have explicitly tested sharks for sex-biased dispersal: the white shark, Carcharadon carcharias , the shortfin mako shark, Isurus oxyrhinchus , and the sandbar shark, Carcharhinus plumbeus , all species that adhere to either coastal or pelagic life histories rather than a combination of the two . In each case, contrasting maternally inherited and biparentally inherited genetic markers indicated dispersive males and philopatric females. The scalloped hammerhead shark, Sphyrna lewini, is a large (up to 420 cm) viviparous (live-bearing) shark with a circumglobal distribution in tropical and warm-temperate waters. S. lewini gives birth to 13–30 pups following an 8–10 month gestation  and is thought to reproduce annually , . Pups are born in shallow coastal nursery habitats where they can be seasonally resident for 3–5 years . Adults are highly migratory and have an unusual coastal-pelagic life history, often schooling over seamounts and near continental and insular shelves to depths possibly in excess of 275 m . Adult S. lewini can occasionally be found in the open ocean, and documented oceanic movements exceed 1500 km . S. lewini populations have declined worldwide as a result of overfishing and bycatch , and are listed as globally endangered by the IUCN Red List of Threatened and Endangered Species .
An earlier analysis of maternal (mtDNA) markers in S. lewini revealed significant genetic structure between oceans and between discontinuous coastlines within oceans . Population structure was low or absent along coastlines and continental margins, a pattern consistent with weak female philopatry to coastal nursery grounds , , . Duncan et al.  proposed that females disperse readily across continuous habitat but rarely across open oceans. With limited sampling and a single matrilineal locus, however, the possibility of male-mediated dispersal could not be evaluated.
An ideal system for untangling male and female components of dispersal would include both maternally- and paternally-inherited markers. Lacking markers for the Y sex chromosomes (males carry the XY karyotype; ), here we rely instead on biparentally-inherited markers to resolve the male contribution to population structure. Nuclear markers such as microsatellites have been shown to useful for estimating population demographics in marine populations , and contrasting nuclear and mitochondrial data have been applied successfully in the past to the identification of differential dispersal patterns between sexes , , , . However few studies to date have addressed a fundamental issue in population genetics that challenges this method, namely the four-fold smaller effective population size (Ne) of the haploid, uniparental mitochondrial DNA (mtDNA) compared to diploid, biparental nuclear DNA (nDNA) , , . These differences in Ne mean that even in the absence of sex-biased migration, mtDNA structure may be greater than nDNA structure due to the differential rate at which the markers attain drift-migration equilibrium.
To illuminate these aspects of Sphyrna lewini reproduction and population structure, we used thirteen biparentally inherited microsatellite markers to genotype 403 sharks collected from eleven locations throughout a global range (Figure 1). Our study addresses two primary issues:
- Sex-biased dispersal: Previous mtDNA analyses by Duncan et al (2006) have ruled out high connectivity across ocean basins for female lineages. Nonetheless the circumglobal range and single-species status of S. lewini indicates high dispersal. Here we contribute results from 13 microsatellite loci to assess the proposal of sex-biased dispersal in S. lewini.
- Population structure and phylogeography: Duncan et al.  proposed an Indo-Pacific origin for S. lewini with subsequent dispersal westward into the Atlantic and eastward into the Central and Eastern Pacific. We address this hypothesis with increased sampling and the application of microsatellite data, which in conjunction with mtDNA data has substantial power to resolve population history.
TEP = Tropical East Pacific, HH = Hawaii, TW = Taiwan, PH = Philippines, EA = East Australia, W = West Australia, SY = Seychelles, SA = South Africa, WAF = West Africa, SC = South Carolina, and GM = Gulf of Mexico. Sample numbers for each site are in parentheses.
Tissue collection and ethics statement
Scalloped hammerhead specimens (fin, muscle, or liver tissue) were acquired in Hawaii by fishing under permit #2008-99 issued by the State of Hawaii's Division of Aquatic Resources to the Hawaii Institute of Marine Biology, and elsewhere were bought from commercial fishermen or purchased from fish markets between 1999 and 2008. Specimens were collected from multiple locations in each of three ocean basins, including (i) Pacific: Tropical East Pacific, Pacific Panama, Hawaii, the Philippines, Taiwan, and Eastern Australia; (ii) Indian: Western Australia, Seychelles, and South Africa; and (iii) Atlantic: Western Africa, Gulf of Mexico, and East Coast USA (South Carolina). Collection sites were grouped together if they were within close geographic proximity and statistical analyses indicated no significant differences between sites; two collections of tissue specimens (N = 23 and N = 20, respectively) were made from proximate locations and grouped to create a single Gulf of Mexico collection. Similarly, two sites in Baja California consisting of 32 and 24 specimens were grouped together with five specimens from Pacific Panama to create one Tropical East Pacific (TEP) sample. When possible, specimens were collected from juvenile sharks (fork length <60 cm) within a proposed nursery area to avoid the confounding effect of sampling adults in feeding areas where distinct breeding populations may overlap .
DNA extraction and microsatellite fragment amplification
Tissue samples (<1 cm3) were stored in 20% dimethylsulfoxide (DMSO) saturated salt (NaCl) buffer  or 75% ethanol (EtOH), and DNA was extracted using a salting-out protocol adapted from Sunnucks and Hales . Individuals were genotyped for thirteen microsatellite loci, including ten species-specific markers  and three markers developed for other species (Table 1) ,  following quality control testing as outlined in Selkoe and Toonen . Unlabeled reverse primers were obtained from Integrated DNA Technologies, Inc. (Coralville, Indiana). Forward primers were labeled with 6-FAM, VIC, NED, and PET proprietary dyes (Applied Biosystems, Foster City, California). PCR reactions consisted of 0.1 U Biolase Taq DNA polymerase (Bioline; Randolph, Massachusetts), 1× Taq buffer, 0.25–0.0625 µM of each primer, 200 µM each dNTP, and 2.0 mM MgCl2. PCR amplification on a MyCycler (Bio-Rad; Hercules, California) consisted of an initial denaturation at 95°C for 4 min, followed by 30 cycles of 1 min at 95°C, 30 s at optimal annealing temperature (Table 1), and 30 s at 72°C, followed by a final extension at 72°C for 20 min. PCR products were resolved with an ABI 3100 automated sequencer and visualized using ABI PRISM GeneMapper Software 4.0 (Applied Biosystems).
We tested for deviation from Hardy–Weinberg Equilibrium (HWE) using Genepop 3.4 , and estimated heterozygosity and tested for linkage disequilibrium using Arlequin 3.11 . Genetic duplicates were detected using the Excel Microsatellite Toolkit . Where significant deviations from Hardy–Weinberg Equilibrium were identified, the program MicroChecker 1  was used to determine whether they were consistent with null alleles, errors due to stutter, large allele dropout, scoring errors, or typography. MicroChecker does this by constructing random genotypes from the alleles observed at each locus, and comparing the distribution of these against the distribution of actual observed genotypes across every allele size class. Probabilities for the number of homozygotes within each size class are calculated using a cumulative binomial distribution . MicroChecker was also used to calculate the frequency of null alleles per locus (r) according to the algorithm implemented in . Allelic richness for each sampling site was estimated using Fstat 220.127.116.11 , which applies a rarefaction method to correct for variation in sample sizes between populations (Table 2). We conducted a regression analysis in jmp 7 (SAS Institute, Cary, NC) to statistically estimate degree of decrease in observed heterozygosity and allelic richness per site with increasing distance from the putative center of origin for S. lewini, the Indo-West Pacific, by using straight-line distance from Indonesia in kilometers as our dependent variable.
Population genetic analyses
We estimated the degree of genetic differentiation among individual sampling sites with pairwise values of FST generated in Arlequin . Interpreting FST values generated from highly variable multiallelic data can be problematic because the maximum FST is constrained by the mean within-subpopulation heterozygosity , . Even in the absence of any shared alleles, highly heterozygous loci will asymptote far below the theoretical maximum (FST = 1), making comparisons between microsatellites and other genetic markers difficult to interpret. To address the downward bias in FST estimates with highly polymorphic loci, we applied the FST standardization approach  as implemented by Meirmans  in RecodeData 0.1. Pairwise FST estimates differ in magnitude but provide the same qualitative patterns and significance values as standardized F′ST estimates (Table 3), so we discuss only the standardized estimates herein.
The program TreeFit  was used to graphically illustrate genetic distances between sampling sites. TreeFit constructs a genetic distance matrix of pairwise FST  and then uses these values to build an evolutionary tree where branch lengths reflect genetic distances between sites. The program then generates an R-squared value that statistically expresses the proportion of variation in the distance matrix that is explained by the tree (i.e., how well the figure represents the data). A high R-squared value indicates that the pairwise FSTs generated in Arlequin are well illustrated by the tree, though it does not provide statistical support for any particular evolutionary model. The evolutionary tree generated in TreeFit was illustrated as a radial dendrogram in TreeView v.1.6.6 .
Global patterns of population subdivision were also examined using Structure 2.3.2 , , which provides an unbiased estimate of the number of gene pools present using a Bayesian model-based clustering method to assign individuals to groups by minimizing linkage disequilbrium and deviations from Hardy-Weinberg expectations. Whereas TreeFit uses pairwise FST to illustrate genetic relationships between sampling sites, Structure's individual-based approach provides a minimum estimate of the number of genetic populations present across S. lewini's range. We employed the admixture model with correlated allele frequencies, as this configuration is appropriate in cases of subtle population structure , and used collection sites as prior information as per . The admixture model allows individuals in the sample to be assigned to single cluster or jointly to two or more clusters if their combined genotypes indicate admixture. Here, the presence of admixed individuals can provide evidence of unsampled population structure. To infer how many clusters or populations (K) are represented in a data set, we plotted the change in the log probability of successive K values (ΔK)  in Jmp 7. ΔK cannot be used for K = 1, but when more than one population is likely, ΔK shows a mode at the actual K more consistently than L(K) and is particularly effective for resolving large microsatellite data sets where individuals may be admixed . We employed a 10,000 step burn-in followed by 10,000 simulations to test K = 1−11 with 10 repetitions each.
Patterns of pairwise differentiation and individual gene pool assignment were used to group collection sites according to the number of populations indicated by Structure. We then assessed the magnitude of genetic differentiation between the K subpopulations in Arlequin with a hierarchical Analysis of Molecular Variance (AMOVA; 20,000 permutations), which partitions genetic variation within sites (FST), among sites within groups (FSC), and among groups (FCT).
We tested for differences between male and female dispersal patterns by comparing pairwise genetic distances between populations for mtDNA and microsatellite DNA (F′ST). F′ST is a standard measure of population differentiation representing the proportion of genetic diversity resulting from allele frequency shifts between sampling sites, with no consideration for genetic distance between individual alleles. For this reason, direct comparisons of F′ST with mitochondrial ΦST, which takes into account both haplotype frequency and genetic distance between haplotypes, can be problematic. We therefore used Arlequin to generate a mitochondrial analogue for pairwise F′ST (herein identified as mtFST) that describes S. lewini population structure solely in terms of haplotype frequencies (Table 4). Paired t tests of means were performed to examine the difference between mtFST, F′ST and ΦST in Jmp 7.
We estimated migration rates (m1 and m2; Table 4) using the coalescent-based program IMa  on the CBSU computing clusters at Cornell University. IMa simulates gene genealogies using MCMC sampling methods, and implements an “isolation with migration” model that does not assume gene flow and genetic drift are in equilibrium. Although IMa2  can handle multiple populations at once, we analyzed all adjacent pairs of populations separately in IMa because IMa2 requires a well-supported phylogeny of the groups of individuals being analyzed . Although some runs were performed on nuclear data alone, MCMC mixing and convergence were poor, so these results are not reported and mtDNA and nDNA data sets were combined for all IMa analyses. We started with an analysis in “MCMC Mode” using the full complement of model parameters (i.e., θ1≠θ2≠θA, and m1≠m2), with broad priors for all, and reduced them in repeated runs to better sample the posterior distribution. Once several replicates converged on the same answer, we recorded the maximum likelihood estimate (MLE) for each parameter and 95% highest posterior density interval or HPD. We converted migration parameters m1 and m2 into the number of migrants per generation (Nm) using the equation Nm = (θ m)/4.
Microsatellite fragments were analyzed for 403 specimens from eleven sampling locations. Micro-Checker  detected no errors resulting from DNA degradation, low DNA concentrations, and primer-site mutations. Microsatellite Toolkit for Excel found no evidence of duplicate or redundant sampling among individuals. Null alleles were detected at low frequencies (r; Table 1) among populations at three loci: Sle018, 054, and 077. These nulls were generally too rare to affect HWE, although Sle077 had a significant P value (P = 0.042) consistent with departure from equilibrium (Table 1). To test whether marginal (though non-significant) differences between expected vs. observed heterozygosities at some loci could confound population level analyses, we removed the four loci that showed that largest difference between Ho and He (Cli-100, Sle018, Sle053, and Sle077) and re-ran pairwise FST. A comparison of pairwise FST values calculated from the subset of 9 loci and from the full data set was non-significant (paired t-test calculated in jmp; t = 1.99, df = 64, P = 0.325). As such, we present results from the full 13 loci here. There was no evidence of linkage disequilibrium among pairs of loci after correction for multiple comparisons.
Observed heterozygosity among geographic sampling sites ranged from Ho = 0.537−0.785, and allelic richness ranged from Ar = 6.91−8.54 with the exception of the Gulf of Mexico (Ar = 4.89; Table 2). Observed heterozygosities were highest in Taiwan, East Australia, and South Africa, while allelic richness was highest in East Australia, West Australia, and the Seychelles (Ar = 7.30−8.49). The East Pacific and the Gulf of Mexico sites had the lowest heterozygosity and allelic richness (Ar = 4.89−6.91). Allelic richness showed significant decrease with increasing distance from Indonesia (R-square = 0.43, P = 0.028), while the decrease of observed heterozygosity was lower and not statistically significant (R-square = 0.25, P = 0.113).
Both corrected (F′ST) and uncorrected (FST) microsatellite based estimates of pairwise population differentiation revealed a general pattern of high contemporary gene flow across ocean basins and along continental margins (average F′ST≈0), with some noteworthy exceptions (Table 3). Significant levels of genetic structure were detected across the 2000 km of open ocean separating Hawaii from the Tropical East Pacific (TEP; F′ST = 0.027, P = 0.001), however no comparisons differentiated Hawaii from the three sampling sites in the West Pacific (average F′ST = −0.063, P = 0.874). No population structure was detected across the Indian Ocean (e.g. between West Australia and the Seychelles, F′ST = −0.010, P = 0.670). Likewise Indian Ocean samples are undifferentiated from Taiwan, Philippines, and Eastern Australia in the West Pacific (average F′ST = −0.018, P = 0.470), indicating contemporary gene flow bridging the Sunda Shelf barrier. South Africa was weakly differentiated from most other Indo-Pacific populations except for the Seychelles, with which it shares a continuous continental shelf (F′ST = −0.036, P = 0.956). While the degree of differentiation indicates some isolation of the South African site, this site shows allele frequencies similar to Hawaii across a distance of 17,600 km (F′ST = −0.018, P = 0.764). Between the Indian and Atlantic Oceans, South Africa is differentiated from West Africa (F′ST = 0.071, P = 0.006). Low but significant structure also differentiates West Africa and South Carolina (F′ST = 0.052, P = 0.042), indicating limited contemporary gene flow across the Atlantic. The Gulf of Mexico was highly differentiated from every other site sampled (average F′ST = 0.438, P<0.001), including the proximate South Carolina site in the West Atlantic (F′ST = 0.201, P<0.001). The highest level of allelic differentiation separates the Gulf of Mexico from the TEP (F′ST = 0.519, P<0.001). The broad mixing of microsatellite alleles across the Indian and Pacific oceans is clearly illustrated in the radial dendrogram of pairwise FST genetic distances (Figure 2), which connects all Indo-Pacific sampling sites with the exception of the TEP (Hawaii, Taiwan, Philippines, East Australia, West Australia, Seychelles, and South Africa) together with short intervening branch lengths. TreeFit assigned an R-squared value of 0.986, indicating that this figure describes the genetic distance reflected in the data set with a high degree of confidence.
Lengths of braches demonstrate allelic similarities between nodes, each of which represents a separate sampling site. R-square value expresses the proportion of variation in the distance matrix that is explained by the dendrogram. TEP = Tropical East Pacific, HH = Hawaii, TW = Taiwan, PH = Philippines, EA = East Australia, W = West Australia, SY = Seychelles, SA = South Africa, WAF = West Africa, SC = South Carolina, and GM = Gulf of Mexico.
Structure  indicates four clusters or populations among our global data set (Figure 3). Following selection of K, we used Structure's individual genotypic assignments to generate pie charts for each sampling site, in addition to using the traditional bar plot to display individual admixture. The pie charts denote the total probability of individuals sampled from that site belonging to each of the four clusters indicated by Structure (C1–C4). Geographical structuring is as follows: Cluster C1 is by far the most widespread, represented in every sampling site from the TEP and Hawaii throughout the Indo-Pacific and Atlantic to South Carolina. Taiwan, Philippines, and South Africa assign primarily to this cluster, though this could be due to low sample size. One small, relatively isolated cluster (C3) is predominantly observed in the Indian Ocean, with high representation in the Seychelles. In South Carolina, about half of the individuals were assigned to C1, the widespread cluster found in the Indo-Pacific, and about half to the predominantly Atlantic C2 cluster shared between West Africa and the Gulf of Mexico. The low level of admixture in South Carolina indicates that samples may have been drawn from a common feeding area where distinct breeding populations may overlap. The Gulf collections were the most homogeneous of any sampling site, with nearly 100% of individuals assigned exclusively to C2. The TEP sample also exhibited high population homogeneity, with the majority of individuals assigned to a single genetic cluster (C4).
Pie charts in the top figure indicate the relative proportion of individuals from each sampling site that assign to each lineage, and the bottom figure shows individual genotypic assignments in a conventional bar plot, organized by sampling site. TEP = Tropical East Pacific, HH = Hawaii, TW = Taiwan, PH = Philippines, EA = East Australia, W = West Australia, SY = Seychelles, SA = South Africa, WAF = West Africa, SC = South Carolina, and GM = Gulf of Mexico.
For the Analysis of Molecular Variance (AMOVA), we grouped sites into four subpopulations as indicated by Structure and tested these for within- and between-group variation. Individuals from the East Pacific and the Gulf of Mexico had the greatest probability of self-assignment and were accordingly analyzed as independent sub-populations. As the remaining nine sites showed extensive admixture, groupings were not immediately obvious and collection sites were therefore clustered by major biogeographic province, these being the Indo-Pacific (central Pacific – West Indian Ocean) and Atlantic . The resulting AMOVA groupings were as follows: (i) Tropical East Pacific; (ii) Indo-Pacific (Hawaii, Taiwan, Philippines, East Australia, South Africa, Seychelles, and West Australia); (iv) Atlantic (South Carolina and West Africa); and (v) Gulf of Mexico. No significant variance was found among sampling sites within groups (FSC = 0.002, P = 0.098). However, 4.88% of the variance was partitioned among groups (FCT = 0.049, P<0.001), indicating that our groupings reflect actual population structure. Global genetic structure was low but significant (FST = 0.035, P<0.001). Though our data indicate gene flow both across the Sunda Shelf and around the tip of South Africa, the presence of a unique multi-locus lineage primarily found within the Indian Ocean (C3, blue in Fig. 3) indicates some isolation in this region, perhaps in the Gulf of Oman.
Genetic distances derived from biparentally-inherited microsatellite DNA (F′ST) and matrilineal mtDNA (mtFST and ΦST) are contrasted in Table 4. Paired Student's t-tests revealed statistically significant differences between F′ST and mtFST in S. lewini (t = −5.025, P<0.001). ΦST and mtFST were highly similar except in a limited number of pairwise comparisons where the magnitude of ΦST was elevated relative to mtFST (though P values remained consistent). This effect was noted particularly in pairwise comparisons that include the Seychelles, West Africa, and West Australia. Because divergent haplotypes coupled with genetic isolation could create a signal of haplotypic distance that increases the magnitude of ΦST relative to mtFST, these data indicate some genetic isolation of Indian Ocean sampling sites relative to the rest of the globe, results similar to those indicated by Structure.
IMa revealed maximum likelihood estimates of the number of migrants per generation (Nm = θm/4) between 0.03 and 7.22. Although 95% posterior probability densities (PPDs) of these estimates were large with several infinite upper boundaries, all PPDs had strong peaks with probabilities falling to zero as m approached zero. We report both Nm1 (estimated migrants per generation from population 2 into population 1) and Nm2 (estimated migrants per generation from population 1 into population 2) for each pair of sites examined, and compare these mixed-marker values to the migration rates for mitochondrial data alone (M) from Duncan et al.  (Table 4). Along continuous coastlines, Nm1/Nm2 were consistently equivalent to or lower than M, while across ocean basins Nm1/Nm2 values were higher than M, sometimes by an order of magnitude.
The distribution of maternal lineages indicates strong restrictions to dispersal between discontinuous coastlines, whereas biparentally-inherited markers reveal much higher connectivity and in some cases nonsignificant population structure across ocean basins. Male-mediated dispersal and gene flow has likely facilitated the contemporary connectivity observed among global Sphyrna lewini populations, while population genetic differences between sites are enhanced by females' ontogenetic requirement for coastal habitat. Populations at the ends of the range in the Tropical East Pacific (TEP) and the Gulf of Mexico are relatively young (<100,000 yr) , whereas the oldest contemporary populations, based on both mtDNA and microsatellite data, are in the central Indo-Pacific .
We identified marked differences between male and female dispersal patterns in Sphyrna lewini, as indicated by contrasting genetic distances derived from matrilineal mtDNA (mtFST and ΦST; ) and biparentally-inherited microsatellite DNA (F′ST; Table 4). As males do not transmit mtDNA haplotypes to subsequent generations, high mtDNA structure and limited or absent microsatellite DNA structure between sites indicates female site fidelity, most likely for the purpose of reproduction.
Overall, F′ST was marginally lower than mtFST and ΦST along continuous coastlines but an order of magnitude lower across ocean basins (Table 4). For example, highly significant mtDNA structure contrasts with a lack of significant structure in microsatellite DNA between the Seychelles in the western Indian Ocean and West Australia (F′ST = −0.009, P = 0.149; mtFST = 0.521, P≤0.001; ΦST = 0.736, P≤0.001), and between Hawaii and all West Pacific sites (Table 4), indicating contemporary male-mediated gene flow across large expanses of open ocean. Male-biased dispersal was also noted across some biogeographic barriers, such as the Tropical East Pacific barrier (between Hawaii and the TEP) and Mid-Atlantic barrier (between East and West Atlantic). Notably, estimates of gene flow across the Atlantic (∼4000 km) were lower than across the Indo-Pacific (∼7000 km), as evidenced by significant genetic structure at microsatellite loci. There was on average an order of magnitude difference in population structure detected by mitochondrial and nuclear DNA in these pairs. Concordantly, IMa indicated that estimates of gene flow among mitochondrial and mixed-marker analysis are similar along continental margins, a finding consistent with generally high gene flow among both males and females along continuous coastlines. Across ocean basins, however, the estimates for mtDNA are far higher than for biparental markers (Table 4), indicating that when dispersal events across oceanic barriers occur, they are largely male-mediated.
Although divergent FST values gleaned from mitochondria are generally expected to be higher and easier to detect than nuclear FST because of mtDNA's haploid nature and uniparental inheritance , the analyses used in this study show marker divergences to be attributable to sex-biased dispersal rather than differences in Ne. First, the large numbers of highly polymorphic markers used in this study lend considerable power to discerning genetic structure at nuclear loci, power that is equal to or greater than mtDNA. In their statistical comparisons between haploid and diploid marker types, Larsson et al.  demonstrated that in situations which include several populations, numerous loci, and many alleles per locus (>5), organelle and nuclear marker types have equal power to detect genetic structure. Second, tests for recent population expansion in Duncan et al. (2006) were consistently non-significant across all sampling sites with the exception of Hawaii. Evidence of stable populations across S. lewini's range decreases the likelihood that the order of magnitude or greater differences between mtDNA and nDNA structure, standardized for marker heterozygosity, are purely an artifact of differences in marker Ne. Further, IMa estimates of gene flow, which are by design robust to differences in Ne, show high migration at mixed markers compared with mtDNA values across ocean basins but not along continuous coastlines, results that are biologically consistent with our hypothesis of male-biased dispersal. Notably, both nuclear DNA markers and gene flow estimates from mixed-marker analysis (Nm) include evidence gleaned from biparental inheritance, yet even with 50% input from matrilines, the differences between marker types clearly signify male-mediated dispersal.
This male-biased dispersal pattern is consistent with the premise that shark reproductive strategies more closely resemble those of sea turtles and marine mammals than other fishes (Musick 1999). It is likely that female reliance on coastal habitat for reproduction can explain the relatively limited vagility of S. lewini females. One caveat to our conclusions is that genetic isolation is not always synonymous with restricted physical movement in sharks. Lemon sharks (Negaprion brevirostris) show philopatry to specific sites , , but may move extensively between breeding seasons. Similarly, contrasting mtDNA and microsatellite inheritance reveal male-biased dispersal between South Africa and Australia in white sharks (C. carcharias; ), however mature females have been shown to migrate between these locations . In nature, males and females may have the same migratory circuit, but female migrations include reproductive philopatry whereas male migrations provide an opportunity for long-distance gene flow. Though mtDNA structure along continental margins is low when compared across ocean basins, both this work and others indicate female site fidelity to coastlines, and perhaps even philopatry to specific coastal embayments , , . The annual reproductive cycle probably precludes long-distance movements by female S. lewini, but telemetry studies would be necessary to link the sex-biased gene flow observed here with reduced transoceanic movement by females.
The statistical parsimony analysis in Duncan et al.  identified basal mtDNA haplotypes in the Indian and West Pacific Oceans, and mitochondrial coalescence analyses found the oldest populations of Sphyrna lewini in the same region (Table 1 and Figure 2 in ). These data further demonstrate that the East Pacific and Gulf of Mexico/West Atlantic sites have high pairwise structure and the lowest population ages, indicating relatively recent origins of contemporary populations . Our dendrogram of microsatellite FST distances (Figure 2) shows that Indo-Pacific sites group together with high allelic similarity, with differential branching into the Tropical East Pacific and Atlantic Oceans. We conclude that the Indo-Pacific, and possibly the West Pacific, is the origin of modern S. lewini, with subsequent divergence eastward into the East Pacific and westward into the Atlantic. These findings are consistent with the IWP Center of Origin hypothesis proposed by Briggs  and advocated by Duncan et al. . Notably, microsatellite allelic richness and admixture is highest within the Indo-Pacific (Ar = 7.30−8.49), and decreases progressively with distance from that region (R-square of Ar = 0.43, P = 0.028). Genetic structure shows the opposite pattern, with lower population structure within the Indo-Pacific and increasing differentiation towards the ends of the global range in the East Pacific and Gulf of Mexico. These sites each show the lowest heterozygosity, admixture, and allelic richness, as well as the highest pairwise differentiation (F′ST = 0.519, P<0.001).
Despite evidence of male migration across thousands of kilometers of open ocean, we identified surprisingly strong genetic differentiation between the Gulf of Mexico and the adjacent South Carolina (West Atlantic) population. Structure shows almost zero admixture in the Gulf population, while the adjacent South Carolina (West Atlantic) site shows a near 1∶1 mix of Gulf (green) and Indo-Pacific (red) types (Figure 3). IMa indicates unidirectional dispersal between these two sites, with estimated migration out of the Gulf of Mexico into the West Atlantic being high relative to other sites (Nm = 7.22), while estimated migration into the Gulf is near zero (Nm = 0.26; Table 4). This pattern is concordant among marker types, indicating restricted gene flow into the Gulf of Mexico for both males and females. However, a similar study on S. lewini mtDNA using samples from adult sharks taken in the shark fin trade  found no such structure separating the Gulf from the West Atlantic. The source of this discrepancy may lie in the fact that Chapman et al.  sampled adults, possibly from a mixed pool of transient migrants. In addition to adults, the current study includes unrelated neonates and juveniles, which are less dispersive and therefore more likely to reflect to the genetic consequences of philopatry.
Reasons for the isolation of the Gulf of Mexico from the proximate West Atlantic in the current study are hard to pinpoint, though it is clear that these samples belong to S. lewini and not the cryptic Sphyrnid species found in the same region . Duncan et al.  proposed that partitioning throughout the West Atlantic, Gulf of Mexico, and Caribbean was largely driven by the distances between nursery sites. Long-distance tagging data on hammerhead sharks are few, and only one S. lewini tagged in the Gulf of Mexico (sex unknown) has been recaptured to date, not far from the catch site , . Similar to the Atlantic, low but significant structure was found between West and South Africa, and between the Philippines, Taiwan, and East Australia (Table 4), possibly reflecting the input of strong female philopatry.
The circumglobal tropical distribution of S. lewini crosses multiple well-documented marine biogeographic barriers and includes seamounts, coralline archipelagos, pelagic blue-water environments, and coastal brackish-water estuaries. In S. lewini, maximum population differentiation between East Pacific and West Atlantic populations is likely due to the difficulty of maintaining circumglobal gene flow, in particular the bridging of large stretches of open water separating the Tropical East Pacific from the Central Pacific, the cold upwelling water around South Africa, and the oceanic gap between East and West Atlantic . Two biogeographic barriers require particular attention here: the Indo-Pacific Barrier and the Isthmus of Panama.
The shallow continental shelf between the Pacific and Indian Oceans is regarded as a substantial barrier to tropical invertebrates and fishes , , , . At glacial maxima, the Sunda Shelf between southeast Asia and Australia/New Guinea is exposed and forms a nearly impenetrable land bridge between Pacific and Indian Oceans. However, the transient nature of this Indo-Pacific Barrier produces different responses by organisms with a variety of dispersal mechanisms. Gaither et al.  noted that most teleost (bony) fishes surveyed across this barrier showed strong population-level or evolutionary separations. However, this does not seem to be the case for the elasmobranchs surveyed to date (Keeney & Heist 2006; Castro et al. 2007), possibly because their high adult vagility allows for connectivity across the Sunda Shelf, even during lowered sea levels. While the scalloped hammerhead shows a population level separation across this barrier with mtDNA, that lack of microsatellite DNA structure over the same area demonstrates male mediated dispersal between the Pacific and Indian Oceans (Table 4).
The Isthmus of Panama separated the East Pacific from the West Atlantic approximately three million years ago , . Mitochondrial coalescence estimates for the TEP and Gulf of Mexico populations show probable range expansion into both areas long after the rise of the Isthmus , indicating that these highly differentiated populations have never been in contact and are linked only through circumglobal gene flow across three ocean basins. In this case, the Isthmus differs from its well-documented role as an emergent vicariant barrier isolating portions of a formerly contiguous population . Instead, we find that S. lewini is prevented from completing its circumglobal range expansion by secondary vicariance due to the Isthmian land barrier. In S. lewini, the secondary vicariant role of the Isthmus of Panama represents a novel and previously undescribed role for the Isthmus.
Notably, such a vicariant speciation event has already marked the evolutionary history of ancestral S. lewini; despite an Indo-Pacific origin, a cryptic sister species was recently discovered in the West Atlantic , a likely product of an earlier Atlantic colonization. The West Atlantic and TEP are also sites of endemism for several coastal members of the genus Sphyrna , indicating that a shared biogeographic mechanism, potentially linked to the presence of the Isthmus of Panama, is driving Sphyrnid diversity in this region.
Conclusions and management implications
The scalloped hammerhead shark (Sphyrna lewini) is a large, mobile predator with a coastal-pelagic life history and global range. Both mtDNA and microsatellite data indicate that modern populations of this tropical species originated in the Indo-West Pacific, then subsequently dispersed into the Atlantic and East Pacific. Although the ability to traverse pelagic habitat indicates high rates of dispersal, female S. lewini display site fidelity to single coastlines, archipelagos, or individual nursery areas and show no evidence of ongoing trans-oceanic movement. Male S. lewini, by contrast, disperse long distances across the open ocean with clear consequences in reproduction and the transmission of gametes, though the frequency of these migrations is unknown. Populations at the ends of the range in the TEP and the Gulf of Mexico have differentiated over time, and are prevented from contact by the Isthmus of Panama. Currently scalloped hammerheads are exposed to massive fishery mortality on a global scale, and their coastal-pelagic life history makes them vulnerable to inshore and estuarine fishing mortality as well as offshore commercial operations. In 2008, the IUCN raised the global status of S. lewini from “threatened” to “endangered” . In the U.S., scalloped hammerheads are grouped with the large coastal species (LCS), the category of sharks that scientists consider most susceptible to commercial overfishing . This group has undergone documented declines throughout the Atlantic, where S. lewini is considered to be a single genetic stock , , a premise that is clearly refuted by population genetic data presented here and in previous work .
In regards to management strategies, the complex population structure observed in S. lewini highlights the need for analyses of genetic markers with multiple lines of inheritance. Single-marker assays using either female or biparentally-inherited loci alone conducted on species with complex reproductive behavior and life history may give a misleading picture of management units , where multiple-locus studies offer more comprehensive results. While mtDNA data from philopatric females may reflect genetic partitioning within ocean basins or along continental margins, arguing for more conservative management units along coastlines, highly dispersive males may be crossing oceans and are potentially being fished far from their location of origin and at both ends of a single migratory circuit. Overall, our evidence of differential dispersal patterns in male and female S. lewini, as well as strong genetic partitioning in the West Atlantic and in other regions of the world, leads us to recommend drastic reevaluation of management units and conservation strategies for this species.
For assisting with global sample collection, we gratefully thank D. Grubbs, M. Shivji, D. Rowat, P. Marko, D. Chapman, D. Itano, Y. Papastamatiou, N. Whitney, J. Chidlow, C. Caraguel, R. Robertson, G. Sedberry, S. Clarke, S. Fennessey, J. Romine, P. Santos, M. Francis, J. Garmin, J. Grady, F. Marquez, A. Amorim, G. Burgess, J. Carlson, G. Naylor, A. Guadiano, E. Heist, F. Magana, J. Schultz, G. Cliff, L. Dagorn, J. Loefer, B. Nishimoto, R. Nu'u, A. Piercy, K. Priede, J. Quattro, W. Robbins, and R. McAuley, as well as the South African Natal Shark Board, Sea World South Africa, Maui Ocean Center, Hilo Division of Aquatic Resources, and the Seychelles Bureau of Standards. For help with genetic analyses we thank J. Eble, S. Daley, M. Blake, A. Milazzo, A. Eggers, and C. Bird. Thanks also to L. Rocha, D. Grubbs, S. Conant, C. Morden, and J. Eble for editorial assistance, and to all the members of the Holland, Toonen, and Bowen Labs. Genetic analyses were made possible by the EPSCoR Evolutionary Genetics Facility at the Hawaii Institute of Marine Biology. This is contribution #1379 from the Hawaii Institute of Marine Biology and contribution #7864 from the School of Ocean and Earth Science and Technology at the University of Hawaii.
Conceived and designed the experiments: TD-E KS RT BB. Performed the experiments: TD-E JC. Analyzed the data: TD-E JC. Contributed reagents/materials/analysis tools: TD-E KS KH HN RT BB. Wrote the paper: TD-E RT BB.
- 1. Neff BD, Pitcher TE (2005) Genetic quality and sexual selection: an integrated framework for good genes and compatible genes. Molecular Ecology 14: 19–38.
- 2. Smith RL (1984) Sperm Competition and the Evolution of Animal Mating Systems. San Diego, CA: Academic Press.
- 3. Zeh DW, Zeh JA (2000) Reproductive mode and speciation: the viviparity-driven conflict hypothesis. Bio Essays 22:
- 4. Favre L, Balloux F, Goudet J, Perrin N (1997) Female-biased dispersal in the monogamous mammal Crocidura russula: evidence from field data and microsatellite patterns. Proceedings of the Royal Society B 264: 127–132.
- 5. Engelhaupt D, Hoelzel AR, Nicholson C, Frantzis A, Mesnick S, et al. (2009) Female philopatry in coastal basins and male dispersion across the North Atlantic in a highly mobile marine species, the sperm whale (Physeter macrocephalus). Molecular Ecology 18: 4193–4205.
- 6. Pusey AE (1987) Sex-biased dispersal and inbreeding avoidance in birds and mammals. Trends in Ecology and Evolution 2: 295–299.
- 7. Musick JA (1999) Ecology and conservation of long-lived marine animals. In: Musick JA, editor. Life in the Slow Lane: Ecology and Conservation of Long-Lived Marine Animals. Bethesda, MD: American Fisheries Society Symposium 23. pp. 1–10.
- 8. Carrier JC, Pratt HL, Castro JI (2004) Reproductive Biology of Elasmobranchs. In: Carrier JC, Musick JA, Heithaus MR, editors. Biology of Sharks and their Relatives. Boca Raton: CRC Press. pp. 269–286.
- 9. Holland KN, Wetherbee BM, Peterson JD, Lowe CG (1993) Movements and distribution of hammerhead shark pups on their natal grounds. Copeia 2: 495–502.
- 10. Branstetter S (1990) Early life-history implications of selected carcharhinoid and lamnoid sharks of the northwest Atlantic. Springfield, VA: National Oceanic and Atmospheric Administration Report 90, National Marine Fisheries Service, U.S. Department of Commerce National Technical Information Service. pp. 17–28.
- 11. Keeney DB, Heist EJ (2006) Worldwide phylogeography of the blacktip shark (Carcharhinus limbatus) inferred from mitochondrial DNA reveals isolation of western Atlantic populations coupled with recent Pacific dispersal. Molecular Ecology 15: 3669–3679.
- 12. Bowen BW, Bass AL, Soares L, Toonen RJ (2005) Conservation implications of complex population structure: lessons from the loggerhead turtle (Caretta caretta). Molecular Ecology 14: 2389–2402.
- 13. Duncan KM, Martin AP, Bowen BW, De Couet HG (2006) Global phylogeography of the scalloped hammerhead (Sphyrna lewini). Molecular Ecology 15: 2239–2251.
- 14. Schrey AW, Heist EJ (2003) Microsatellite analysis of population structure in the shortfin mako (Isurus oxyrhinchus). Canadian Journal of Fisheries and Aquatic Sciences 60: 670–675.
- 15. Hoelzel AR, Shivji MS, Magnussen J, Francis MP (2006) Low worldwide genetic diversity in the basking shark (Cetorhinus maximus). Biology Letters 2: 639–642.
- 16. Pardini AT, Jones CS, Noble LR, Kreiser B, Malcolm H, et al. (2001) Sex-biased dispersal of great white sharks. Nature 412: 139–140.
- 17. Portnoy DS, McDowell JR, Heist EJ, Musick JA, Graves JE (2010) World phylogeography and male-mediated gene flow in the sandbar shark, Carcharhinus plumbeus. Molecular Ecology 19: 1994–2010.
- 18. NMFS (2008) Final Amendment 2 to the Consolidated Atlantic Highly Migratory Species Fishery Management Plan Silver Spring, MD: National Oceanic and Atmospheric Administration, National Marine Fisheries Service, U.S. Department of Commerce Highly Migratory Species Management Division.
- 19. White WT, Bartron C, Potter IC (2008) Catch composition and reproductive biology of Sphyrna lewini (Griffith & Smith) (Carcharhiniformes, Sphyrnidae) in Indonesian waters. Journal of Fish Biology 72: 1675–1689.
- 20. Hazin F, Fischer A, Broadhurst M (2001) Aspects of reproductive biology of the scalloped hammerhead shark, Sphyrna lewini, off northeastern Brazil. Environmental Biology of Fishes 61: 151–159.
- 21. Duncan KM, Holland KN (2006) Habitat use, growth rates and dispersal patterns of juvenile scalloped hammerhead sharks (Sphyrna lewini) in a nursery habitat. Marine Ecology Progress Series 312: 211–221.
- 22. Compagno LJV, Dando M, Fowler SL (2005) Sharks of the World. Princeton, NJ: Princeton University Press.
- 23. Kohler NE, Turner PA (2001) Shark tagging: a review of conventional methods and studies. Environmental Biology of Fishes 60: 191–223.
- 24. IUCN (2008) The IUCN Red List of Threatened Species: An Analysis of Fish on the 2008 IUCN Red List 〈www.iucnredlist.org/fish〉. Conservation International, Arizona State University, Texas A&M University, University of Rome, University of Virginia, Zoological Society London.
- 25. DiBattista JD, Feldheim KA, Thibert-Plainte X, Gruber SH, Hendry AP (2008) A genetic assessment of polyandry and breeding site fidelity in lemon sharks. Molecular Ecology 17: 3337–3351.
- 26. Chapman DD, Babcock EA, Gruber SH, DiBattista JD, Franks BR, et al. (2009) Long-term natal site-fidelity by immature lemon sharks (Negaprion brevirostris) at a subtropical island. Molecular Ecology 18: 3500–3507.
- 27. Maddock MB, Schwartz FJ (1996) Elasmobranch cytogenetics: methods and sex chromosomes. Bulletin of Marine Science 58: 147–155.
- 28. Tallmon DA, Gregovitch D, Waples RS, Baker CS, Jackson J, et al. (2010) When are genetic methods useful for estimating contemporary abundance and detecting population trends? Molecular Ecology Resources 10: 684–692.
- 29. Kerth G, Mayer F, Petit E (2002) Extreme sex-biased dispersal in the communally breeding, nonmigratory Bechstein's bat (Myotis bechsteinii). Molecular Ecology 11: 1491–1498.
- 30. Karl SA, Castro ALF, Lopez JA, Charvet P, Burgess GH (2011) Phylogeography and conservation of the bull shark (Carcharhinus leucas) inferred from mitochondrial and microsatellite DNA. Conservation Genetics 12: 371–382.
- 31. Birky CW, Maruyama T, Fuerst P (1983) An approach to population and evolutionary genetic theory for genes in mitochondrial and chloroplasts and some results. Genetics 103: 513–527.
- 32. Larsson LC, Charlier J, Laikre L, Ryman N (2009) Statistical power for detecting genetic divergence- organelle versus nuclear markers. Conservation Genetics 10: 1255–1264.
- 33. Sunnucks P (2000) Efficient genetic markers for population biology. Trends in Ecology & Evolution 15: 199–203.
- 34. Seutin G, White BN, Boag PT (1991) Preservation of avian blood and tissue samples for DNA analyses. Canadian Journal of Zoology 69: 82–90.
- 35. Sunnucks P, Hales DF (1996) Numerous transposed sequences of mitochondrial cytochrome oxidase I–II in aphids of the genus Sitobion (Hemiptera: Aphididae). Molecular Biology and Evolution 13: 510–524.
- 36. Nance HA, Daly-Engel TS, Marko PB (2009) Microsatellite loci for the endangered scalloped hammerhead shark, Sphyrna lewini. Molecular Ecology Resources 9: 955–957.
- 37. Chapman DD, Prodohl PA, Gelsleicher J, Manire CA, Mahmood SS (2004) Predominance of genetic monogamy by females in a hammerhead shark, Sphyrna tiburo: implications for shark conservation. Molecular Ecology 13: 1965–1974.
- 38. Keeney DB, Heist EJ (2003) Characterization of microsatellite loci isolated from the blacktip shark and their utility in requiem and hammerhead sharks. Molecular Ecology Notes 3: 501–504.
- 39. Selkoe KA, Toonen RJ (2006) Microsatellites for ecologists: a practical guide for using and evaluating microsatellite markers. Ecology Letters 9: 615–629.
- 40. Raymond M, Rousset F (1995) GENEPOP Version 1.2. Population genetics software for exact tests and ecumenicism. Journal of Heredity 86: 248–249.
- 41. Excoffier L, Laval LG, Schneider S (2005) Arlequin ver. 3.0: An integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1: 47–50.
- 42. Park SDE (2001) Trypanotolerance in West African Cattle and the Population Genetic Effects of Selection (PhD Thesis): University of Dublin.
- 43. van Oosterhout CV, Hutchinson W, Wills D, Shipley P (2004) MICRO-CHECKER: software for identifying and correcting genotype errors in microsatellite data. Molecular Ecology Notes 4: 535–538.
- 44. Weir BS (1996) Genetic Data Analysis II. Sunderland, Massachusetts: Sinauer Associates.
- 45. Goudet J (1995) FSTAT (version 1.2): a computer program to calculate F-statistics. Journal of Heredity 86: 485–486.
- 46. Hedrick P (2005) A standardized genetic differentiation measure. Evolution 59: 1633–1638.
- 47. Meirmans PG (2006) Using the AMOVA framework to estimate a standardized genetic differentiation measure. Evolution 60: 2399–2402.
- 48. Kalinowski ST (2009) How well do evolutionary trees describe genetic relationships among populations? Heredity 102: 506–513.
- 49. Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population structure. Evolution 38: 1358–1370.
- 50. Page RDM (1996) TREEVIEW: An application to display phylogenetic trees on personal computers. Computer Applications in the Biosciences 12: 357–358.
- 51. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945–959.
- 52. Hubisz MJ, Falush D, Stephens M, Pritchard JK (2009) Inferring weak population structure with the assistance of sample group information. Molecular Ecology Resources 9: 1322–1332.
- 53. Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14: 2611–2620.
- 54. Hey J, Nielsen R (2007) Integration with the Felsenstein equation for improved Markov chain Monte Carlo methods in population genetics. Proceedings of the National Academy of Sciences 104: 2785–2790.
- 55. Hey J (2010) Isolation with migration models for more than two populations. Molecular Biology and Evolution 27: 905–920.
- 56. Briggs JC (1974) Marine zoogeography. New York: McGraw-Hill.
- 57. DiBattista JD, Feldheim KA, Gruber SH, Hendry AP (2008) Are indirect genetic benefits associated with polyandry? Testing predictions in a natural population of lemon sharks. Molecular Ecology 17: 783–795.
- 58. Bonfil R, Meyer M, Scholl MC, Johnson R, O'Brien S, et al. (2005) Transoceanic migration, spatial dynamics, and population linkages of white sharks. Science 310: 100–103.
- 59. Chapman DD, Pinhal D, Shivji MS (2009) Tracking the fin trade: genetic stock identification in western Atlantic scalloped hammerhead sharks Sphyrna lewini. Endangered Species Research 9: 221–228.
- 60. Briggs JC (1992) The marine East Indies: centre of origin? Global Ecology and Biogeography Letters 2: 149–156.
- 61. Quattro JM, Stoner DS, Driggers WB, Anderson CA, Priede KA, et al. (2006) Genetic evidence of cryptic speciation within hammerhead sharks (Genus Sphyrna). Marine Biology 148: 1143–1155.
- 62. Rocha LA, Craig MT, Bowen BW (2007) Phylogeography and the conservation genetics of coral reef fishes. Coral Reefs 26: 501–512.
- 63. Barber PH, Erdmann MV, Palumbi SR (2006) Comparative phylogeography of three codistributed stomatopods: origins and timing of regional diversification in the coral triangle. Evolution 60: 1825–1839.
- 64. Craig MT, Eble JA, Robertson DR, Bowen BW (2007) High genetic connectivity across the Indian and Pacific Oceans in the reef fish Myripristis berndti (Holocentridae). Marine Ecology Progress Series 334: 245–254.
- 65. Lacson JM, Clark S (1995) Genetic divergence of Maldivian and Micronesian denies of the damselfishes Stegastes nigricans, Chrysiptera biocellata, C. glauca and C. leucopoma (Pomacentridae). Marine Biology 121: 585–590.
- 66. Planes S, Fauvelot C (2002) Isolation by distance and vicariance drive genetic structure of a coral reef fish in the Pacific Ocean. Evolution 56: 378–399.
- 67. Gaither MR, Toonen RJ, Robertson DR, Planes S, Bowen BW (2010) Genetic evaluation of marine biogeographical barriers: perspectives from two widespread Indo-Pacific snappers (Lutjanus kasmira and Lutjanus fulvus). Journal of Biogeography 37: 133–147.
- 68. Briggs JC (2006) Proximate sources of marine biodiversity. Journal of Biogeography 33: 1–10.
- 69. Coates AG, Collins LS, Aubrey MP, Berggren WA (2004) The geology of the Darien, Panama, and the late Miocene-Pliocene collision of the Panama arc with northwestern South America. Geological Society of America Bulletin 116: 1327–1344.
- 70. NMFS (2003) Final amendment 1 to the fishery management plan for Atlantic tunas, swordfish, and sharks. Silver Spring, MD: National Oceanic and Atmospheric Administration, National Marine Fisheries Service, U.S. Department of Commerce Highly Migratory Species Management Division.
- 71. Piercy AN, Carlson JK, Sulikowski JA, Burgess GH (2006) Age and growth of the scalloped hammerhead shark, Sphyrna lewini, in the north-west Atlantic Ocean and Gulf of Mexico. Marine and Freshwater Research 58: 34–40.