Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Phylogeographic Analyses of Submesophotic Snappers Etelis coruscans and Etelis “marshi” (Family Lutjanidae) Reveal Concordant Genetic Structure across the Hawaiian Archipelago

  • Kimberly R. Andrews ,

    Affiliations Hawai'i Institute of Marine Biology, University of Hawai'i, Kane'ohe, Hawaii, United States of America, School of Biological & Biomedical Sciences, Durham University, South Road, United Kingdom

  • Virginia N. Moriwake,

    Affiliations Hawai'i Institute of Marine Biology, University of Hawai'i, Kane'ohe, Hawaii, United States of America, Department of Oceanography, University of Hawai'i, Honolulu, Hawaii, United States of America

  • Christie Wilcox,

    Affiliations Hawai'i Institute of Marine Biology, University of Hawai'i, Kane'ohe, Hawaii, United States of America, Cell and Molecular Biology Graduate Program, University of Hawai'i, Honolulu, Hawaii, United States of America

  • E. Gordon Grau,

    Affiliation Hawai'i Institute of Marine Biology, University of Hawai'i, Kane'ohe, Hawaii, United States of America

  • Christopher Kelley,

    Affiliations Hawai'i Institute of Marine Biology, University of Hawai'i, Kane'ohe, Hawaii, United States of America, Hawai'i Undersea Research Lab, University of Hawai'i, Honolulu, Hawaii, United States of America

  • Richard L. Pyle,

    Affiliation Bernice P. Bishop Museum, Honolulu, Hawaii, United States of America

  • Brian W. Bowen

    Affiliation Hawai'i Institute of Marine Biology, University of Hawai'i, Kane'ohe, Hawaii, United States of America

Phylogeographic Analyses of Submesophotic Snappers Etelis coruscans and Etelis “marshi” (Family Lutjanidae) Reveal Concordant Genetic Structure across the Hawaiian Archipelago

  • Kimberly R. Andrews, 
  • Virginia N. Moriwake, 
  • Christie Wilcox, 
  • E. Gordon Grau, 
  • Christopher Kelley, 
  • Richard L. Pyle, 
  • Brian W. Bowen


The Hawaiian Archipelago has become a natural laboratory for understanding genetic connectivity in marine organisms as a result of the large number of population genetics studies that have been conducted across this island chain for a wide taxonomic range of organisms. However, population genetic studies have been conducted for only two species occurring in the mesophotic or submesophotic zones (30+m) in this archipelago. To gain a greater understanding of genetic connectivity in these deepwater habitats, we investigated the genetic structure of two submesophotic fish species (occurring ∼200–360 m) in this archipelago. We surveyed 16 locations across the archipelago for submesophotic snappers Etelis coruscans (N = 787) and E. “marshi” (formerly E. carbunculus; N = 770) with 436–490 bp of mtDNA cytochrome b and 10–11 microsatellite loci. Phylogeographic analyses reveal no geographic structuring of mtDNA lineages and recent coalescence times that are typical of shallow reef fauna. Population genetic analyses reveal no overall structure across most of the archipelago, a pattern also typical of dispersive shallow fishes. However some sites in the mid-archipelago (Raita Bank to French Frigate Shoals) had significant population differentiation. This pattern of no structure between ends of the Hawaiian range, and significant structure in the middle, was previously observed in a submesophotic snapper (Pristipomoides filamentosus) and a submesophotic grouper (Hyporthodus quernus). Three of these four species also have elevated genetic diversity in the mid-archipelago. Biophysical larval dispersal models from previous studies indicate that this elevated diversity may result from larval supplement from Johnston Atoll, ∼800 km southwest of Hawaii. In this case the boundaries of stocks for fishery management cannot be defined simply in terms of geography, and fishery management in Hawaii may need to incorporate external larval supply into management plans.


Understanding the genetic connectivity of marine populations can provide valuable information about evolutionary mechanisms in the marine environment, as well as effective management and conservation strategies. The Hawaiian Archipelago has become a natural laboratory for understanding genetic connectivity of marine organisms, including more than 30 marine species across a wide taxonomic range (reviewed in [1]). A comparison of these species reveals at least four geographically concordant population genetic barriers, providing evidence for some commonality in population genetic structure despite the wide variety of life history characters across species. However, underlying this pattern is high variability in population structure among species, with no one species demonstrating all four of these barriers, some dispersive species showing no barriers, and many species having additional barriers. Furthermore, species with similar life history characters often do not share the same genetic barriers, indicating that any single species cannot be used as an exemplar for dispersal in the Hawaiian Archipelago.

The population genetic studies conducted thus far in the Hawaiian Archipelago have included only two species which occur at the lower margin of photosynthetic activity in the mesophotic and submesophotic zones, the Hawaiian Grouper, called “Hapu'upu'u” in Hawai'i (Hyporthodus quernus, previously Epinephelus quernus) [2], [3] and the Crimson Jobfish, called “Opakapaka” in Hawai'i (Pristipomoides filamentosus) [4]. These studies found interesting patterns of genetic structure: both species showed evidence of high connectivity across the Hawaiian Archipelago, but both species also showed evidence for weak but significant genetic divergence at several sites in the middle of the archipelago. The Hawaiian grouper also had elevated genetic diversity at a mitochondrial DNA (mtDNA) marker in the mid-archipelago. Bio-physical simulations of larval dispersal indicated that this genetic distinctiveness of the mid-archipelago may result from connectivity between the mid-archipelago and Johnston Atoll, the closest landmass to the Hawaiian Archipelago (Figure 1) [3], [5]. These simulations also indicated that directional dispersal from the outer edges of the archipelago toward the mid-archipelago may contribute to the elevated mtDNA diversity [3].

Figure 1. Map of the Hawaiian Archipelago and Johnston Atoll showing sampling locations and geographic division between the Northwestern Hawaiian Islands and the Main Hawaiian Islands.

The pattern of genetic structure and high genetic diversity in the mid-archipelago has not been observed in other fish species studied in Hawai'i thus far. These studies include 15 shallow-water reef fish species, most of which exhibit high connectivity across most or all of the archipelago, with the exception of one damselfish (Dascyllus albisella) that shows strong genetic divergence among all islands sampled across the archipelago, and one surgeonfish (Zebrasoma flavescens) that shows low genetic structure [6][13]. The pattern of genetic structure and high diversity in the mid-archipelago has not been observed for invertebrate species either (e.g. [14], [15]).

To test the hypothesis that the pattern of genetic structure and high diversity in the mid-archipelago is consistent for other deepwater fishes, we conducted population genetic analyses of two submesophotic fishes which occur across the archipelago: the Flame Snapper, called “Onaga” in Hawai'i (Etelis coruscans) and the Ruby Snapper, called “Ehu” in Hawai'i (Etelis “marshi”, currently being revised from Etelis carbunculus, Andrews et al. in prep). Adult E. coruscans and E. marshi typically occur at deeper depths (200–320 m for E. coruscans; 200–360 m for E. marshi) than adult P. filamentosus and H. quernus (80–240 m for P. filamentosus, and 120–280 m for H. quernus; but adult H. quernus occur at <10 m depth in the far northwestern atolls of the Hawaiian Archipelago) [16][18]. However, juveniles of these species are sometimes found at shallower, mesophotic depths [16], [19], [20].

The adult habitat of all four of these species is similar in that it consists of hard substrate with high structural complexity and/or high relief [21]. Despite their similar habitat, however, these species have different habitat use, and these differences may influence adult dispersal and genetic connectivity. Adult E. marshi and H. quernus are more closely associated with the bottom substrate including crevasses and under ledges (C.K. pers. obs., [16]) and therefore may exhibit greater genetic structure than other species due to lower dispersal by juveniles and adults. The close association with the bottom for these species may be due to partitioning of feeding niches between benthos and water column, a greater need for protection due to small adult size (E. marshi), or a suspected territorial, haremic social structure (H. quernus). In contrast, E. coruscans and P. filamentosus are more pelagic feeders that usually occur 2–50 meters above the bottom.

To investigate the population genetic structure of E. coruscans and E. marshi, we used mtDNA cytochrome b (cytb) sequences and 10–11 microsatellite loci from 16 sample sites across the archipelago (Figure 1, Table 1, Table 2). Our study was initially motivated by fishery management concerns for these popular food fishes subject to commercial and recreational catches. However, our study is also one of the first to examine the population genetic structure of a submesophotic species, and thus we aimed to explore the influence of this unique habitat on the evolution of fishes compared to shallow-water habitats. For example, deepwater habitat occupied by submesophotic fishes may have lower water movement than shallow habitat due to the lack of Ekman surface transport and waves, which could result in less dispersive larvae and stronger genetic structure. We tested this hypothesis through qualitative comparisons of the population genetic structure of E. coruscans and E. marshi with the population genetic structure reported in previous studies for two other submesophotic fishes (P. filamentosus and H. quernus) and shallow-water fishes in the Hawaiian Archipelago. Additionally, submesophotic habitats may lie below the zone disrupted by sea level changes during glaciations, and thus communities in this deepwater habitat may be older and more stable than their shallow counterparts. We tested this hypothesis by estimating mtDNA coalescent times for E. coruscans and E. marshi, and comparing these values with previously published mtDNA coalescent times from shallow-water fishes.

Table 1. Sample sizes, diversity indices, and sampling years for mtDNA cytochrome b (cytb) and microsatellite loci across sample sites in the Hawaiian Archipelago for Etelis coruscans. h, haplotype diversity; π, nucleotide diversity; AR allele richness; Ho, observed heterozygosity; He, expected heterozygosity.

Table 2. Sample sizes, diversity indices, and sampling years for mtDNA cytochrome b (cytb) and microsatellite loci across sample sites in the Hawaiian Archipelago for Etelis marshi. h, haplotype diversity; π, nucleotide diversity; AR allele richness; Ho, observed heterozygosity; He, expected heterozygosity.


Specimens consisted of fin or muscle tissue collected throughout the Hawaiian Archipelago from 16 locations each for E. coruscans (N = 787) (Table 1, Figure 1) and E. marshi (N = 770) (Table 2, Figure 1). Specimens were collected by commercial fishers who recorded GPS coordinates for each sampling location and stored the specimens frozen or in salt-saturated DMSO buffer [22]. Most fishing activity for these species takes place at a depth of 200–300 m, with distances from shore ranging from a few hundred meters to as far as 65 km. Specimens were collected in 1997–1998 or 2003–2007 (Table 1, Table 2). Genomic DNA was extracted using a phenol chloroform method [23], DNeasy extraction kits (Qiagen, Valencia, CA, USA), or the Hotshot method [24].

MtDNA sequencing

PCRs amplified portions of mtDNA cytochrome b gene for both species. Primers used for E. coruscans were Cyb-05 L15020 (GCCAACGGCGCATCCTTCTTCTT [25]) and Cyb-07 H15573 (AATAGGAAGTATCATTCGGGTTTGATG [26]), amplifying approximately 560 bp. Primers used for E. marshi were designed for this study: EhucybF (TCAGTCGCACACATCTGCCG) and EhucybR (AGTGCAACAAGGACGGCTGC) and amplified 524 bp. PCRs were performed in 15 µl volumes containing 1× MangoMix (Bioline, Taunton MA, USA) and 0.2 µM each primer. For E. coruscans, cycle conditions were as follows: 94°C for 1 min; 35 cycles of 94°C for 30 sec, annealing temperature 55°C for 45 sec; and 72°C for 30 sec; and a final 72°C extension for 10 min. Cycle conditions for E. marshi were the same with the exception of the annealing temperature step, which was 68°C for 30 sec. To clean PCR products, 7.5 units of Exonuclease I and 0.75 units of FastAP alkaline phosphatase (Fermentas Life Sciences, Ontario, Canada) were added to 7.5 µl of PCR product, and aliquots were incubated at 37°C for 60 minutes and 85°C for 15 minutes. PCR products were then sequenced in one direction with an ABI3730 automated sequencer (Applied Biosystems, Foster City, CA). Questionable or low quality sequences were resequenced in the forward direction. Sequences were edited and aligned using GENEIOUS PRO 5.6.2 (Biomatters, LTD, Auckland, NZ).

Microsatellite genotyping

A total of 16 variable microsatellite loci were developed using genomic DNA extracted from E. coruscans and E. marshi through a procedure including restriction enzyme digestion, microsatellite enrichment, cloning of DNA fragments, and DNA sequencing of clones (sensu [27]) (Table 3). Five of these loci were PCR-amplified for both species. An additional five loci were amplified for E. coruscans, resulting in a total of 10 loci for this species; and an additional six loci were amplified for E. marshi, resulting in a total of 11 loci for this species (Table 3). Multiplex PCRs were carried out using fluorescent dye-labeled forward primers and Qiagen Type-It Microsatellite PCR Kits, with three or four loci included per multiplex reaction, and following reaction conditions recommended by Qiagen. PCR products were separated on ABI 3730XL or ABI 3130XL genetic analyzers, with all PCR products from each primer set run exclusively on only one of these two analyzers to avoid bias in fragment size assignments. Fragment sizes were scored using GENEMAPPER 4.0 (Applied Biosystems).

Table 3. Characteristics of 16 microsatellite loci developed for Etelis coruscans and Etelis marshi, including number of alleles (k), observed heterozygosity (Ho), and expected heterozygosity (He) across the Hawaiian Archipelago for each locus.

MSTOOLS 3.1 [28] was used to identify identical genotypes to confirm that no individual fish specimens were present more than once in the dataset. Each microsatellite locus was tested for departures from Hardy Weinberg equilibrium and linkage equilibrium using ARLEQUIN 3.11 [29]. Each locus was also tested for null alleles using MICRO-CHECKER [30]. Null allele frequencies were estimated using FREENA [31].

Genetic Diversity and Population Structure

For the mtDNA, nucleotide (π) and haplotype (h) diversities were obtained with ARLEQUIN. The nucleotide substitution models used to calculate genetic distance were Tamura Nei+gamma = 0.403 [32] for E. coruscans and Tamura Nei for E. marshi; these were the best-fit models chosen using the AIC method in jModeltest 0.1.1 [33]. For microsatellites, observed heterozygosity, expected heterozygosity, allele richness, and total number of alleles were calculated for each locus using FSTAT [34] and ARLEQUIN.

For mtDNA, genetic similarity between geographic locations was investigated with median-joining haplotype networks for both E. coruscans and E. marshi using NETWORK [35]. The maximum parsimony (MP) option was used to remove unnecessary median vectors and links [35].

Genetic similarity between geographic locations was further investigated by calculating population pairwise ΦST values for mtDNA and FST for microsatellites with ARLEQUIN. For mtDNA analyses, genetic distance was calculated using nucleotide substitution models as described above. For both genetic assays, significance of pairwise values was tested using 20,000 permutations. For microsatellites, the influence of each individual locus on multi-locus pairwise FST values was investigated by removing one locus at a time; and the influence of null alleles on FST analyses was investigated by estimating global FST values with and without correcting for null alleles using the ENA method as implemented in FREENA [31].

To investigate the influence of temporal sampling on the estimation of allele frequencies and inference of genetic structure, we conducted pairwise FST analyses comparing samples collected more than three years apart within a given geographic location. These analyses were conducted for any geographic sample location containing more than 10 individuals each per temporal period (i.e. Lisianski and O'ahu for E. coruscans; and Northampton, Necker, and Ni'ihau for E. marshi).

Population genetic structure was further analyzed using Bayesian clustering analyses implemented in STRUCTURE 2.3.3 [36] for microsatellite loci, using the admixture and correlated allele frequency models. The burn-in length was set at 105 steps, followed by 106 steps. Analyses were run five times for each of K = 1 to K = 16 to test for consistency of estimates of P(X|K), where K is the number of clusters. Bayesian clustering analyses were conducted both with and without use of sampling location as a prior; using sampling location as a prior has been shown to recover population structure at lower levels of divergence, without bias towards assigning structure when it is not present [37].

Relationships between geographic distance and genetic divergence were investigated using GENEPOP 4.0.10 [38], [39] for both mtDNA and microsatellite data. Mantel tests (10,000 permutations) and Spearman Rank correlation tests were used to examine correlations between geographic distance vs. genetic divergence (ΦST/(1−ΦST) [40]). Geographic distance was calculated as the shortest great-circle distance between the approximate centers of sampling sites.

To resolve evolutionary histories, Tajima's D and Fu's Fs tests [41] for departure from mutation-drift equilibrium were conducted with ARLEQUIN using mtDNA cytb sequences. Large negative values of Tajima's D and Fu's Fs are expected to occur if populations have experienced selection or recent expansions [42][44]. For Tajima's D, a significant positive value is expected if populations are admixed or experiencing diversifying selection. Fu's Fs is expected to have greater power than Tajima's D for detecting population expansions [45]. To estimate the age, historic female effective size, and post-expansion female effective size of Hawaiian populations, a coalescence analysis was performed with ARLEQUIN, assuming a cytb rate of 2% per million years between lineages (1% within lineages) calibrated with other marine fishes [46], [47] Generation time was estimated at 10 years based on available life history data: recent evidence indicates that female E. marshi in the Hawaiian Archipelago may mature at two years of age (E. DeMartini, pers. comm.), and the maximum age for E. coruscans and E. marshi in the Hawaiian Archipelago is estimated at 13 years [48] (although this may be an underestimate of the maximum age given that E. carbunculus in the Indian Ocean have lifespans of at least 35 years [49]). Because both mutation rate and generation time are not validated for Etelis species, corresponding estimates of effective population size and population age should be interpreted with caution. However, they should suffice to provide first order approximations.


Microsatellite allele fragment lengths and cytb GenBank Accession No.'s (KF920464 to KF920552) are reported for each individual specimen in Table S1 for E. coruscans and in Table S2 for E. marshi.

Identification of matching microsatellite and mtDNA genotypes indicated that two E. marshi specimens were present twice in the dataset: one specimen was present twice in the Maro Reef sample, and the other present twice in the Necker sample. One of each duplicate was removed from the dataset for subsequent mtDNA and microsatellite analyses. No E. coruscans specimens were present more than once in the dataset.

Microsatellite quality control

For E. coruscans, three loci showed evidence of deviation from HWE after Bonferroni correction, including locus EtelisO22O2 (seven geographic locations), EtelisG23O1 (four locations), and EtelisC19O2 (one location). These loci also exhibited an excess of homozygotes at two or more locations (Etelis O22O2 – 11 locations, EtelisG23O1 – 11 locations, Etelis C19O2 – 2 locations), indicating that deviations from HWE may result from null alleles. Three additional loci exhibited evidence of homozygote excess at one or more locations: EtelisI11O2 (five locations), EtelisM15E1 (one location), and EtelisG8E1 (one location). FREENA indicated a relatively low frequency of null alleles across loci and populations (range: 0.00–0.25, mean: 0.03).

For E. marshi, only one locus (EtelisG8E1) showed evidence of deviation from HWE after Bonferroni correction; this locus deviated from HWE at 12 geographic locations. This locus also exhibited an excess of homozygotes at 11 locations, indicating that null alleles may be responsible for deviations from HWE. Two additional loci exhibited evidence for homozygote excess: EtelisA14E2 (six locations) and EtelisE20O1 (one location). Loci which exhibited evidence for homozygote excess generally had higher null allele frequency estimates across populations (EtelisG8E1 range: 0.13–0.32, mean: 0.22; EtelisA14E2: 0.00–0.23, mean: 0.10; EtelisE20O1: 0.00–0.20, mean: 0.01) than other loci (0.00–0.18, mean: 0.01).

No loci showed evidence for linkage disequilibrium for either E. coruscans or E. marshi after Bonferroni correction. Global FST values calculated with and without correcting for null alleles had overlapping 95% confidence intervals for both species. Additionally, removing one locus at a time had little impact on pairwise FST results for either species. For E. coruscans, average deviations between all pairwise FST values calculated using all loci versus all except one locus ranged from 0.0021 (when removing EtelisE2E2) to 0.0059 (when removing EtelisE20O2). For E. marshi, average deviations between all pairwise FST values calculated using all loci versus all except one locus ranged from 0.0015 (when removing EtelisA14E2) to 0.0133 (when removing EtelisE6E1). For each species, the greatest deviations in average FST occurred with the removal of a locus that showed no evidence of null alleles. Below we report results from analyses using all loci.

Genetic diversity and population structure

Editing of cytb DNA sequence data resulted in a 490 bp fragment for E. coruscans and a 436 bp fragment for E. marshi. Diversity values (h and π) were consistently higher for E. coruscans than E. marshi. For E. coruscans, h ranged from 0.67 to 0.84 and π ranged from 0.0019–0.0039 (Table 1), with the highest values for both h and π occurring at St. Rogatien in the middle of the archipelago (Figure 1). For E. marshi, two locations (Salmon Bank and Lisianski) had zero diversity (h = π = 0); however, this low diversity may be related to the small sample sizes at these locations (n = 11 and n = 12) (Table 2). For E. marshi at all other geographic locations, h ranged from 0.15 to 0.49 and π ranged from 0.00043 to 0.00185 (Table 2), with the highest value for h occurring at Gardner and for π occurring at Brooks Banks, and the second highest value for π occurring at Gardner. Higher diversity values for E. coruscans than E. marshi were not driven by the longer cytb fragment resolved for E. coruscans; when diversity values were calculated using only the region of overlap between cytb sequences for the two species (379 bp), E. coruscans still had consistently higher diversity values than E. marshi, except at Gardner, where haplotype diversity was slightly higher for E. marshi (data not shown).

For microsatellites, observed heterozygosity across the Hawaiian Archipelago for E. coruscans ranged from 0.643 to 0.739, and expected heterozygosity ranged from 0.726 to 0.813 (Table 1). For E. marshi, observed heterozygosity ranged from 0.552 to 0.631, and expected heterozygosity ranged from 0.604 to 0.664 (Table 2).

For E. coruscans, NETWORK analyses resulted in multiple maximum parsimony networks, all of which were similar in structure, with one of these reported here (Figure 2a). For E. marshi, NETWORK analyses produced only one maximum parsimony network (Figure 2b). Networks for both species showed overall low genetic divergence between haplotypes, with no more than 5 mutations separating any two haplotypes. For both E. coruscans and E. marshi, haplotype networks revealed no obvious segregation of haplotypes by geographic location.

Figure 2. Median-joining network of cytochrome b haplotypes for (a) Etelis coruscans and (b) Etelis marshi obtained from the program NETWORK (Bandelt et al. 1999).

Each circle represents a haplotype; circle sizes are proportional to the frequency of haplotypes; and line lengths are proportional to the number of mutational steps between haplotype sequences.

For E. coruscans, pairwise ΦST analyses for mtDNA revealed significant divergence for 20 out of 120 pairwise comparisons, with all except one of these comparisons involving locations in the mid-archipelago (Raita through Brooks Banks) (Table 4). For microsatellite loci, only three pairwise FST comparisons revealed significant divergence (Maro Reef. vs. Molokai, Twin Banks vs. Molokai, and Nihoa vs. Kahoolawe) (Table 4). For Bayesian clustering analyses without using sampling location as a prior, average posterior probabilities were consistently highest at K = 1. When sampling location was used as a prior, average posterior probabilities were consistently highest at K = 3; however, visual inspection of the output from all runs of K = 3 indicated only two consistent population clusters: (1) Kaua'i and (2) all other locations (Figure 3a).

Figure 3. Bayesian clustering analysis results obtained with the program STRUCTURE 2.3.3 (Pritchard 2000) using sample location as a prior for (a) Etelis coruscans, K = 2 and (b) Etelis marshi, K = 3.

Table 4. Pairwise population genetic divergence for mtDNA cytb sequences (ΦST, below diagonal) and microsatellites (FST, above diagonal) across the Hawaiian Archipelago for Etelis coruscans.

For E. marshi, pairwise ΦST analyses for mtDNA revealed significant divergence for 9 out of 120 pairwise comparisons, and each of these comparisons involved Gardner (adjacent to St. Rogatien) (Table 5). Pairwise FST comparisons for microsatellite loci revealed significant divergence for 4 out of 120 pairwise comparisons, with all but one of these comparisons (Northampton vs. Niihau) involving a location in the mid-archipelago (Gardner or Raita) (Table 5). For Bayesian clustering analyses that did not use sampling location as a prior, average posterior probabilities were consistently highest at K = 1. When sampling location was used as a prior, average posterior probabilities were consistently highest at K = 2, and visual inspection of the output indicated two consistent population clusters: (1) St. Rogatien in the middle of the archipelago, and (2) all other locations (Figure 3b).

Table 5. Pairwise population genetic divergence for mtDNA cytb sequences (ΦST, below diagonal) and microsatellites (FST, above diagonal) across the Hawaiian Archipelago for Etelis marshi.

For both species, analyses comparing samples collected during different time periods within a given geographic location resulted in no significant pairwise ΦST or FST values for any location or any genetic marker (P>0.05, data not shown), indicating that temporal sampling had little impact on the estimation of allele frequencies for those locations. For all other locations, all specimens were collected exclusively (or almost exclusively) from one time period (i.e. within three years of each other). The finding of little genetic structure across the archipelago for all markers across samples collected during different time periods suggests that temporal sampling had little impact on allele frequency estimates.

For E. coruscans, genetic divergence was significantly correlated with geographic distance for both mtDNA (P = 0.044) and microsatellites (P = 0.042), although the correlation coefficient (r) was low for both marker types (mtDNA r = 0.126; microsatellites r = 0.171) (Figure 4). For E. marshi, genetic divergence was not significantly correlated with geographic distance for mtDNA (P = 1.000) or microsatellites (P = 0.833).

Figure 4. Correlation between genetic distance (pairwise ΦST or FST) and geographic distance between sample sites for (a) mtDNA cytochrome b (P = 0.044, r = 0.126) and (b) microsatellites (P = 0.042, r = 0.171) for Etelis coruscans.

Correlations between genetic and geographic distance were non-significant for both marker types for Etelis marshi.

All values of Fu's Fs and most values of Tajima's D were negative and significant (P<0.05) for all sites for both species; locations which did not have a significant Tajima's D included Pearl & Hermes, Twin Banks, and Kaho'olawe for E. coruscans (Table 6), and Gardner for E. marshi (Table 7). Coalescence analyses of the mtDNA data provided estimates for age of Hawaiian populations ranging from 5,100 to 9,180 years for E. coruscans (Table 6); and ranging from 344,000 to 378,000 years for all locations for E. marshi except Gardner, which had an age estimate of 8,030 (Table 7). Initial female effective population sizes ranged from 0 to 35.9 for E. coruscans and were 0 at all sites for E. marshi; these low values indicate a very small population size prior to population expansion. For E. coruscans, current female effective population sizes at all locations reached the maximum possible value in ARLEQUIN, which was 99,999 (reported in Table 6 as ∞); for E. marshi, current female effective population sizes ranged from 2,000 to 7,850 for all locations except Gardner, which had the maximum possible value in ARLEQUIN.

Table 6. Tajima's D values, Fu's FS values, and mismatch distribution parameter estimates calculated using mtDNA cytb sequences for Etelis coruscans. t: coalescent time; ⊖0 (Nft = 0): initial female effective population size; ⊖1(Nft = 1): post-expansion female effective population size.

Table 7. Tajima's D values, Fu's FS values, and mismatch distribution parameter estimates calculated using mtDNA cytb sequences for Etelis marshi. t: coalescent time; ⊖0 (Nft = 0): initial female effective population size; ⊖1(Nft = 1): post-expansion female effective population size.


Population genetic analyses across the Hawaiian Archipelago for the deepwater fishes E. coruscans and E. marshi revealed similar patterns of genetic structure, and this structure is concordant with the submesophotic fishes surveyed thus far in Hawai'i, H. quernus and P. filamentosus [2], [3], [27]. Each of these four species exhibited little or no genetic structure across the archipelago, with the exception of genetically divergent populations occurring in the center of the archipelago (e.g., Gardner, St. Rogatien, Brooks Banks, Necker). Additionally, the highest mtDNA diversity occurred at a location in the mid-archipelago for three of these four species (at Gardner and Brooks Banks for H. quernus and E. marshi; at St. Rogatien for E. coruscans). E. coruscans was the only species with evidence of a genetically divergent population outside of the mid-archipelago (Kaua'i).

Coalescence analyses of mtDNA data indicate that the Hawaiian population of E. coruscans is much younger than sympatric E. marshi (<10,000 years vs >300,000 years; Table 6, Table 7). This may be due to a recent bottleneck in E. coruscans. A more intriguing possibility, that this species is a recent arrival to Hawaii, awaits testing with specimens from outside the archipelago. Current female effective population sizes for E. coruscans reached the maximum value in ARLEQUIN, >99,999 (reported in Table 6 as ∞) as compared to <10,000 for E. marshi. Hence coalescence analyses, while offering only first order approximations of these population parameters, indicate that the E. coruscans population in Hawaii is younger and larger than E. marshi over recent evolutionary history. The exception to both trends is E. marshi at Gardner in the mid-archipelago, which had an age estimate of only 8,030 (Table 7), and an estimated female effective population size of the maximum possible value in ARLEQUIN, >99,999. We provisionally attribute this finding to violations of coalescence assumptions at a location receiving larval input from an unsampled, genetically distinct location outside the Hawaiian Archipelago (see below), although this ad hoc interpretation is subject to further scrutiny.

Life history traits that might influence the population structure of these four species (spawning time, larval duration, adult vagility, etc.) are similar but not identical across species. The most dispersive life stage for these fishes is probably the larval stage. Spawning timing may influence dispersal of pelagic eggs and larvae due to seasonality of oceanographic currents in the Hawaiian Archipelago [5]. H. quernus, P. filamentosus, E. coruscans, and E. marshi are all broadcast spawners, and all have similar spawning periods (approximately six months), although P. filamentosus has a slightly longer spawning period (approximately ten months) [16]. Peak spawning activity occurs during the summer for each of these species except H. quernus, which peaks in the spring.

The length of time that larvae are in the water column (the pelagic larval duration, or PLD) also likely influences dispersal and genetic connectivity for these four deepwater fish [50]. Eggs of H. quernus, P. filamentosus, E. coruscans, and E. marshi remain in the water column no more than 48 hours before hatching [16]. Biophysical modeling indicates that these eggs could travel up to 50 km (A. Vaz unpublished data). The length of time the larvae remain in the water column after hatching is not well known. For H. quernus, the PLD is estimated to be 35–45 days (R. Nichols and E. DeMartini pers. comm.). PLD is not known for P. filamentosus, E. coruscans, or E. marshi, but eteline lutjanids typically have longer PLDs than H. quernus [51], [52]. For P. filamentosus, a captive rearing study suggests a potential PLD estimate of 52–120 days, based on observations of larvae associating with the bottom as early as 52 days and the completion of larval metamorphosis by 120 days after hatching [16].

Fishes that are benthic as adults (E. marshi and H. quernus) might be expected to have more structure than fishes that are epibenthic as adults (E. coruscans and P. filamentosus). Contrary to this expectation, neither of the bottom-dwelling species showed higher structure in pairwise FST comparisons. The significant structure at mid-archipelago (based on microsatellites) was FST = 0.008–0.010 and FST 0.008–0.013 for the bottom dwellers, and FST = 0.011–0.013 and FST = 0.004–0.066 for the pelagic roamers. Further, significant isolation by distance was detected in one bottom dweller (H. quernus) and one pelagic feeder (E. coruscans), providing no clear pattern based on habitat preference. Despite modest differences in spawning time, PLD, and habitat use, Hawaiian submesophotic fishes have concordant population structure that is not shared with fishes that inhabit adjacent shallow habitat.

Connectivity with Johnston Atoll

In previous genetic studies of submesophotic reef fishes (H. quernus and P. filamentosus) in the Hawaiian Archipelago, the authors hypothesized that the pattern of population structure and high diversity in the mid-archipelago is driven by connectivity with Johnston Atoll [2][4]. Johnston Atoll is the nearest landmass to the Hawaiian Archipelago, with the closest Hawaiian island/atoll being French Frigate Shoals, in the mid-archipelago between Brooks Banks and Necker, 865 km north-northeast of Johnston (Figure 1). Several studies have found connectivity between Johnston and the Hawaiian Archipelago for a broad taxonomic range of fishes and invertebrates (e.g., [8], [11], [14], [15], [53]). In some cases, these species exhibit greater genetic connectivity between Johnston and the mid-archipelago than between Johnston and other parts of the archipelago (e.g., [14], [15], [53]). In addition, biogeographic studies support a dispersal corridor between Johnston and the mid-archipelago. Based on marine faunal affinities, Johnston is regarded as part of the Hawaiian biogeographic province [54][59]. Finally, a biophysical model of oceanic dispersal indicates two dispersal corridors from Johnston Atoll to the Hawaiian Archipelago for marine species with PLDs greater than 40 days, one centered at French Frigate Shoals, and the other at Kaua'i [5]. The latter is notable because Kaua'i is the only location that was strongly differentiated in STRUCTURE analysis of E. coruscans. In addition, biophysical modeling of larval dispersal for the submesophotic fish H. quernus indicated connectivity between Johnston and Gardner, but not other locations in Hawai'i [3].

Unfortunately none of the genetic studies on the submesophotic reef fishes of Hawai'i have included specimens from Johnston Atoll, and therefore the hypothesis of genetic connectivity between Hawai'i and Johnston has not been directly tested. The PLDs of E. coruscans and E. marshi are likely greater than 45 days, which is sufficient to maintain connectivity between the mid-archipelago and Johnston Atoll according to the biophysical models [5]. Additionally, each of these species has a wide-spread distribution across the Indo-Pacific, indicating high dispersal capabilities. Studies are currently underway to examine population genetic structure across the ranges of these species to get a greater understanding of dispersal patterns.

Comparing genetic structure revealed by different markers and analysis types

Although the pattern of genetic structure and high diversity in the mid-archipelago was consistent across the four submesophotic reef fishes in the Hawaiian Archipelago, not all genetic markers or analyses consistently revealed these patterns within species. For example, mtDNA analyses of H. quernus, E. coruscans, and E. marshi revealed a greater number of significant pairwise divergence values than did microsatellite analyses; and P. filamentosus exhibited the opposite pattern. Obtaining different results for mtDNA and microsatellite FST analyses is common, and is usually attributed to differences in mutation rates, modes of inheritance, and/or levels of power between marker types [60][64]. Obtaining different results for different types of analyses that use the same marker is common as well (reviewed in [60], [61]), and this was observed in some cases for the fishes studied here. For example, microsatellite Bayesian clustering analyses sometimes indicated genetically divergent populations that were not detected with microsatellite pairwise FST analyses, and vice versa (e.g. Bayesian clustering analyses indicated Kaua'i was divergent from other locations for E. coruscans, but pairwise FST analyses did not). These subtle differences in genetic structure for different markers and analyses and the overall pattern of weak genetic structure across the archipelago for all four species are consistent with a scenario in which migration across the archipelago is around the level at which genetic methods lose power to detect restrictions in gene flow. Genetic methods lose power at levels of gene flow that are still low enough to maintain demographically independent populations; this is especially true when effective population sizes are large, as in many marine fish and invertebrate populations [65], [66]. The idea that demographically relevant population structure occurs within the Hawaiian Archipelago is also supported by the presence of significant genetic isolation by distance for two of the deepwater species (H. quernus, [67]; and E. coruscans). Therefore the structure detected for deepwater fish using genetic methods is likely biologically relevant on an ecological timescale.

Submesophotic population structure

The potential influence of unique oceanographic features of the submesophotic habitat on population genetic structure is largely unknown. Species inhabiting these environments may be less dispersive if larvae are not subject to the high-energy environment of surface waters [68]. A previous study of a submesophotic snapper (Pristipimoides multidens) in Indonesia and northern Australia supported this idea, with surprisingly strong genetic structure found over short geographic distances (e.g. genetically divergent groups separated by as little as 191 km) [69]. In contrast, the population structure observed in the four submesophotic fishes is typical of shallow-water members of the same taxonomic families [10], [70] with the exception of the mid-archipelago structure that so far is unique to submesophotic fishes. Indeed, it is not certain that spawned eggs or larvae remain at submesophotic depths, and so this expectation may not be valid. Another possibility is that deeper communities may be less susceptible to the disruption that shallow communities experience during sea level drops associated with glaciations [68], [71], [72]. During the last glacial maximum, sea levels dropped at least 130 m below contemporary levels [73]. Collectively the four fishes evaluated here range down to ∼330 m, and adult Etelis specimens have been documented below that depth; this would seem to be ample buffer against disturbance in the upper 130 m. Furthermore, the primary habitat of these species consists of rocky substrate, not living coral, and therefore the abundance of suitable habitat may not be substantially influenced by sea level changes. Nevertheless, these fish show the short coalescence times that are typical of shallow reef populations ([10], [70] for examples): ∼6,000 years for E. coruscans and ∼344,000 years for E. marshi. While this topic requires further investigation, especially for the coral reef fauna at mesophotic depths, the evidence here indicates that submesophotic fishes experience the same type of episodic crashes as confamilial species in shallow depths.

Conservation & Management

E. coruscans, E. marshi, H. quernus, and P. filamentosus comprise the majority of a valuable deepwater fishery in Hawai'i [74][76] and they have experienced overfishing in recent years [48]. The Northwestern Hawaiian Islands (Nihoa to Kure) have been off-limits to the fishery since 2010 due to the designation of this region as the Papahānaumokuākea Marine National Monument. However, the presence of fishing pressure in the Main Hawaiian Islands necessitates an understanding of stock structure across the archipelago for the effective management of this fishery. Population genetic analyses of these four deepwater fishes point to the mid-archipelago as a genetically diverse region possibly with larval supplement from an external source. Genetic analyses reveal high levels of connectivity across the rest of the archipelago, with the possible exception of a genetically divergent population of E. coruscans at Kaua'i. The genetic data provide several indications that dispersal between Hawaiian Islands may be at a level low enough to create some level of demographic independence, but high enough to prevent genetic divergence between these regions. Other methods for measuring demographic connectivity, such as tagging studies and bio-physical modeling of larval dispersal, will be an important complement to these genetic data for the assessment of stock structure.

In the realm of dispersive and migratory marine species, there are many cases in which genetically divergent groups are distributed in complex geographic configurations, and these cases often necessitate complex management strategies. For example, both sea turtles and salmon have genetically differentiated stocks at spawning habitats, but these stocks are diffuse and thoroughly mixed in oceanic feeding habitats [77], [78]. For marine species with pelagic larvae, an increasing number of studies are finding evidence that complex genetic structure depends more on oceanographic currents or habitat distribution than on simple geographic distance [79], [80]. The Hawaiian submesophotic fishes also appear to have geographically complex population genetic structure that is not defined solely by geographic distance, because locations in the mid-archipelago are genetically distinct from nearby locations. If future studies find this genetic divergence to be driven by dispersal from Johnston Atoll, then larval supplement to the local Hawaiian stock from Johnston Atoll may need to be incorporated into management strategies.


Population genetic studies have revealed high variability in patterns of genetic structure across the Hawaiian Archipelago for shallow marine species, even for species with similar life history characters. In contrast, all four submesophotic fishes surveyed to date showed concordant patterns of genetic structure, with dispersal across the entire archipelago, yet genetically divergent populations occurring in the mid-archipelago. Genetic divergence at the mid-archipelago may be driven by connectivity with Johnston Atoll, located outside of the Hawaiian Archipelago. Therefore management strategies may need to consider the impact of larval supply from outside the Hawaiian Archipelago, a supplement that has yet to be quantified. Our results also indicate that submesophotic populations may be as dispersive as their shallow water counterparts, and may have been equally unstable over geological time.

Supporting Information

Table S1.

Specimen ID, sampling location, microsatellite allele lengths (bp), and cytb GenBank Accession numbers for Etelis coruscans specimens. Missing data coded by “0”.


Table S2.

Specimen ID, sampling location, microsatellite allele lengths (bp), and cytb GenBank Accession numbers for Etelis marshi specimens. Missing data coded by “0”.



Mahalo to the fishermen in the Hawaiian bottomfish fishery for their assistance in sample collection throughout the Hawaiian Islands. For facilitating this project we thank Sam Pooley, Michael Seki, and Robert Moffitt at the NOAA Pacific Islands Fisheries Science Center, Kitty Simonds and Mark Mitsuyasu at the Western Pacific Regional Fishery Management Council, Aulani Wilhelm and Randall Kosaki at the Papahānaumokuākea Marine National Monument, and the Hawai'i State Division of Aquatic Resources. Thanks to Jonathan Puritz, Jeff Eble, Joseph DiBattista, Matthew Craig, Luiz Rocha, Iria Fernandez-Silva, Joshua Copus, Derek Skillings, Zoltan Szabo, and especially Robert Toonen and Steve Karl for their intellectual input. We thank the staff of the HIMB EPSCoR Evolutionary Genetics Core Facility and especially Amy Eggers for assistance with genotyping, and the University of Hawai'i Advanced Studies of Genomics, Proteomics and Bioinformatics sequencing facility for their assistance with DNA sequencing and genotyping. Thanks to the staff at Hawai'i Institute of Marine Biology, and especially director Jo-Ann Leong, for their support throughout this project. We thank Robert Humphreys, Edward DeMartini, Robert Moffitt, and Ana Vaz for comments that substantially improved the manuscript. This is contribution #1581 from the Hawai'i Institute of Marine Biology, #9078 from the School of Ocean and Earth Science and Technology, and UNIHI-SEAGRANT-JC-05-53 from the University of Hawai'i Sea Grant Program.

Author Contributions

Conceived and designed the experiments: KRA VNM EGG CK RLP BWB. Performed the experiments: KRA VNM CW. Analyzed the data: KRA VNM CW. Contributed reagents/materials/analysis tools: EGG CK BWB. Wrote the paper: KRA BWB.


  1. 1. Toonen RJ, Andrews KR, Baums IB, Bird CE, Concepcion GT, et al. (2011) Defining boundaries for ecosystem-based management: A multispecies case study of marine connectivity across the Hawaiian Archipelago. Journal of Marine Biology Article ID 460173.
  2. 2. Rivera MAJ, Kelley CD, Roderick GK (2004) Subtle population genetic structure in the Hawaiian grouper, Epinephelus quernus (Serranidae) as revealed by mitochondrial DNA analyses. Biological Journal of the Linnean Society 81: 449–468.
  3. 3. Rivera MAJ, Andrews KR, Kobayashi D, Wren J, Kelley C, et al. (2011) Genetic analyses and simulations of larval dispersal reveal distinct populations and directional connectivity across the range of the Hawaiian Grouper (Epinephelus quernus). Journal of Marine Biology Article ID 765353.
  4. 4. Gaither MR, Jones SA, Kelley C, Newman SJ, Sorenson L, et al. (2011) High connectivity in the deepwater snapper Pristipomoides filamentosus (Lutjanidae) across the Indo-Pacific with Isolation of the Hawaiian Archipelago. Plos One 6: e28913
  5. 5. Kobayashi DR (2006) Colonization of the Hawaiian Archipelago via Johnston Atoll: a characterization of oceanographic transport corridors for pelagic larvae using computer simulation. Coral Reefs 25: 407–417.
  6. 6. Craig MT, Eble JA, Bowen BW, Robertson DR (2007) High genetic connectivity across the Indian and Pacific Oceans in the reef fish Myripristis berndti (Holocentridae). Marine Ecology Progress Series 334: 245–254.
  7. 7. Ramon ML, Nelson PA, De Martini E, Walsh WJ, Bernardi G (2008) Phylogeography, historical demography, and the role of post-settlement ecology in two Hawaiian damselfish species. Marine Biology 153: 1207–1217.
  8. 8. Eble JA, Toonen RJ, Bowen BW (2009) Endemism and dispersal: comparative phylogeography of three surgeonfishes across the Hawaiian Archipelago. Marine Biology 156: 689–698.
  9. 9. Craig MT, Eble JA, Bowen BW (2010) Origins, ages and population histories: comparative phylogeography of endemic Hawaiian butterflyfishes (genus Chaetodon). Journal of Biogeography 37: 2125–2136.
  10. 10. Gaither MR, Bowen BW, Toonen RJ, Planes S, Messmer V, et al. (2010) Genetic consequences of introducing allopatric lineages of Bluestriped Snapper (Lutjanus kasmira) to Hawaii. Molecular Ecology 19: 1107–1121.
  11. 11. DiBattista JD, Wilcox C, Craig MT, Rocha LA, Bowen BW (2011) Phylogeography of the Pacific Blueline Surgeonfish Acanthurus nigroris reveals a cryptic species in the Hawaiian Archipelago. Journal of Marine Biology Article ID 839134.
  12. 12. Eble JA, Rocha LA, Craig MT, Bowen B (2011) Not all larvae stay close to home: Insights into marine population connectivity with a focus on the Brown Surgeonfish (Acanthurus nigrofuscus). Journal of Marine Biology Article ID 518516.
  13. 13. Reece JS, Bowen BW, Smith DG, Larson A (2011) Comparative phylogeography of four Indo-Pacific moray eel species (Muraenidae) reveals comparable ocean-wide genetic connectivity despite five-fold differences in available adult habitat. Marine Ecology Progress Series 437: 269–277.
  14. 14. Timmers MA, Toonen RJ, Andrews K, deMaintenton MJ, Brainard R (2011) Widespread dispersal of the crown-of-thorns sea star, Acanthaster planci, across the Hawaiian Archipelago and Johnston Atoll. Journal of Marine Biology Article ID 934269.
  15. 15. Skillings D, Bird CE, Toonen RJ (2011) Gateways to Hawai'i: Genetic population structure of the tropical sea cucumber Holothuria atra. Journal of Marine Biology Article ID 783030.
  16. 16. WPRFMC (2012) Final fishery management plan for coral reef ecosystems of the western Pacific region, volume III, Essential Fish Habitat for Management Unit Species, Appendix 3, Essential Fish Habitat descriptions, Part 1: Hawaiian bottomfish.
  17. 17. WPRFMC (1988) Amendment 2 to the fishery management plan for the bottomfish and seamount groundfish fisheries of the Western Pacific Region. Western Pacific Regional Fishery Management Council, Honolulu.
  18. 18. DeMartini EE, Friedlander AM (2004) Spatial patterns of endemism in shallow-water reef fish populations of the Northwestern Hawaiian Islands. Marine Ecology-Progress Series 271: 281–296.
  19. 19. Moffitt RB, Parrish FA (1996) Habitat and life history of juvenile Hawaiian Pink Snapper, Pristipomoides filamentosus. Pacific Science 50: 371–381.
  20. 20. DeMartini EE, Lau BB (1999) Morphometric criteria for estimating sexual maturity in two snappers, Etelis carbunculus and Pristipomoides sieboldii. Fishery Bulletin 97: 449–458.
  21. 21. Ralston S (1979) A description of the bottomfish fisheries of Hawaii, American Samoa, Guam and the Northern Marianas. Honolulu: Western Pacific Regional Fisheries Management Council.
  22. 22. Seutin G, White BN, Boag PT (1991) Preservation of avian blood and tissue samples for DNA analyses. Canadian Journal of Zoology 69: 82–90.
  23. 23. Cummings SA, Thorgaard GH (1994) Extraction of DNA from fish blood and sperm. Biotechniques 17: 426–430.
  24. 24. Meeker ND, Hutchinson SA, Ho L, Trecle NS (2007) Method for isolation of PCR-ready genomic DNA from zebrafish tissues. Biotechniques 43: 610–614.
  25. 25. Meyer A (1993) Molecular phylogenetic studies of fishes. In: Beaumont AR, editor. Evolution and Genetics of Aquatic Organisms. London: Chapman and Hall.
  26. 26. Taberlet P, Meyer A, Bouvet J (1992) Unusual mitochondrial DNA polymorphism in two local populations of blue tit Parus caeruleus. Molecular Ecology 1: 27–36.
  27. 27. Gaither MR, Toonen RJ, Sorenson L, Bowen BW (2010) Isolation and characterization of microsatellite markers for the Crimson Jobfish, Pristipomoides filamentosus (Lutjanidae). Conservation Genetics Resources 2: 169–172.
  28. 28. Park SDE (2001) Trypanotolerance in West African Cattle and the Population Genetic Effects of Selection [PhD thesis]. Dublin: University of Dublin.
  29. 29. 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.
  30. 30. Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P (2004) MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes 4: 535–538.
  31. 31. Chapuis MP, Estoup A (2007) Microsatellite null alleles and estimation of population differentiation. Molecular Biology and Evolution 24: 621–631.
  32. 32. Tamura K, Nei M (1993) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution 10: 512–526.
  33. 33. Posada D (2008) jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution 25: 1253–1256.
  34. 34. Goudet J (2001) FSTAT, a program to estimate and test gene diversities and fixation indices, Accessed: 2014 Feb 17.
  35. 35. Bandelt HJ, Forster P, Rohl A (1999) Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16: 37–48.
  36. 36. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945–959.
  37. 37. 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.
  38. 38. Raymond M, Rousset F (1995) GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity 86: 248–249.
  39. 39. Rousset F (2008) Genepop'007: a complete reimplementation of the Genepop software for Windows and Linux. Molecular Ecology Resources 8: 103–106.
  40. 40. Rousset F (1997) Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145: 1219–1228.
  41. 41. Tajima F (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585–595.
  42. 42. Aris-Brosou S, Excoffier L (1996) The impact of population expansion and mutation rate heterogeneity on DNA sequence polymorphism. Molecular Biology and Evolution 13: 494–504.
  43. 43. Tajima F (1989) The effect of change in population size on DNA polymorphism. Genetics 123: 597–601.
  44. 44. Fu YX (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147: 915–925.
  45. 45. Ramos-Onsins SE, Rozas J (2002) Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution 19: 2092–2100.
  46. 46. Bowen BW, Bass AL, Rocha LA, Grant WS, Robertson DR (2001) Phylogeography of the trumpetfishes (Aulostomus): Ring species complex on a global scale. Evolution 55: 1029–1039.
  47. 47. Reece JS, Bowen BW, Smith DG, Larson A (2010) Molecular phylogenetics of moray eels (Muraenidae) demonstrates multiple origins of a shell-crushing jaw (Gymnomuraena, Echidna) and multiple colonizations of the Atlantic Ocean. Molecular Phylogenetics and Evolution 57: 829–835.
  48. 48. Brodziak J, Courtney D, Wagatsuma L (2011) Stock assessment of the main Hawaiian Islands Deep7 bottomfish complex through 2010. U.S. Dept. of Commerce, NOAA Technical Memorandum NOAA-TM-NMFS-PIFSC-29.
  49. 49. Andrews AH, Kalish JM, Newman SJ, Johnston JM (2011) Bomb radiocarbon dating of three important reef-fish species using Indo-Pacific Δ14C chronologies. Marine and Freshwater Research 62: 1259–1269.
  50. 50. Selkoe KA, Toonen RJ (2011) Marine connectivity: a new look at pelagic larval duration and genetic metrics of dispersal. Marine Ecology-Progress Series 436: 291–305.
  51. 51. Leis J (1987) Review of the early life history of tropical groupers (Serranidae) and snappers (Lutjanidae). In: Polovina J, Ralston S, editors. Tropical snappers and groupers: biology and fisheries management. Boulder, CO: Westview Pr. pp. 189–237.
  52. 52. Leis JM, Lee K (1994) Larval development in the lutjanid subfamily Etelinae (Pisces): The genera Aphareus, Aprion, Etelis and Pristipomoides. Bulletin of Marine Science 55: 46–125.
  53. 53. Leray M, Beldade R, Holbrook SJ, Schmitt RJ, Planes S, et al. (2010) Allopatric divergence and speciation in coral reef fish: the three-spot Dascyllus, Dascyllus trimaculatus, species complex. Evolution 64: 1218–1230.
  54. 54. Briggs JC, Bowen BW (2012) A realignment of marine biogeographic provinces with particular reference to fish distributions. Journal of Biogeography 39: 12–30.
  55. 55. Grigg RW (1981) Acropora in Hawaii USA 2. Zoogeography. Pacific Science 35: 15–24.
  56. 56. Grigg RW, Wells JW, Wallace C (1981) Acropora in Hawaii USA 1. History of the scientific record, systematics and ecology. Pacific Science 35: 1–13.
  57. 57. Maragos JE, Jokiel PL (1986) Reef corals of Johnston Atoll: one of the world's most isolated reefs. Coral Reefs 4: 141–150.
  58. 58. Mundy BC (2005) Checklist of the fishes of the Hawaiian archipelago. Bishop Museum Press, Honolulu, Hawaii.
  59. 59. Randall JE (2007) Reef and shore fishes of the Hawaiian Islands: Sea Grant College Program, University of Hawaii, Honolulu, Hawaii. 546 p.
  60. 60. Waples RS, Gaggiotti O (2006) What is a population? An empirical evaluation of some genetic methods for identifying the number of gene pools and their degree of connectivity. Molecular Ecology 15: 1419–1439.
  61. 61. Ryman N, Palm S, Andre C, Carvalho GR, Dahlgren TG, et al. (2006) Power for detecting genetic divergence: differences between statistical methods and marker loci. Molecular Ecology 15: 2031–2045.
  62. 62. Avise JC (2004) Molecular Markers, Natural History, and Evolution. Sunderland, MA: Sinauer Associates, Inc.
  63. 63. DiBattista JD, Rocha LA, Craig MT, Feldheim KA, Bowen BW (2012) Phylogeography of two closely related Indo-Pacific butterflyfishes reveals divergent evolutionary histories and discordant results from mtDNA and microsatellites. Journal of Heredity 103: 617–629.
  64. 64. Karl SA, Toonen RJ, Grant WS, Bowen BW (2012) Common misconceptions in molecular ecology: Echos of the modern synthesis. Molecular Ecology 21: 4171–4189.
  65. 65. Waples RS (1998) Separating the wheat from the chaff: patterns of genetic differentiation in high gene flow species. Conservation Biology 89: 438–450.
  66. 66. Waples RS, Punt AE, Cope JM (2008) Integrating genetic data into management of marine resources: how can we do it better? Fish and Fisheries 9: 423–449.
  67. 67. Rivera MAJ (2002) Population Genetic Structure of the Commercially Important Hawaiian Grouper Epinephelus quernus (Serranidae) using Mitochondrial and Nuclear Microsatellite DNA: PhD thesis, University of California, Berkeley, 145 pp.
  68. 68. Lesser MP, Slattery M, Leichter JJ (2009) Ecology of mesophotic coral reefs. Journal of Experimental Marine Biology and Ecology 375: 1–8.
  69. 69. Ovenden JR, Salini J, O'Connor S, Street R (2004) Pronounced genetic population structure in a potentially vagile fish species (Pristipomoides multidens, Teleostei; Perciformes; Lutjanidae) from the East Indies triangle. Molecular Ecology 13: 1991–1999.
  70. 70. Gaither MR, Bowen BW, Bordenave TR, Rocha LA, Newman SJ, et al. (2011) Phylogeography of the reef fish Cephalopholis argus (Epinephelidae) indicates Pleistocene isolation across the Indo-Pacific barrier with contemporary overlap in the coral triangle. BMC Evolutionary Biology 11: 189
  71. 71. Fauvelot C, Bernardi G, Planes S (2003) Reductions in the mitochondrial DNA diversity of coral reef fish provide evidence of population bottlenecks resulting from Holocene sea-level change. Evolution 57: 1571–1583.
  72. 72. Ludt WB, Bernal MA, Bowen BW, Rocha LA (2012) Living in the past: Phylogeography and population histories of Indo-Pacific wrasses (Genus Halichoeres) in shallow lagoons versus outer reef slopes. PLoS One 7: e38042
  73. 73. Voris HK (2000) Maps of Pleistocene sea levels in Southeast Asia: shorelines, river systems and time durations. Journal of Biogeography 27: 1153–1167.
  74. 74. Pooley SG (1987) Demand considerations in fisheries management – Hawaii's market for bottom fish. In: Polovian JJ, Ralston S, editors. Tropical Snappers and Groupers: Biology and Fisheries Management. Boulder, CO: Westview Press. pp. 605–638.
  75. 75. Hospital J, Pan M (2009) Demand for Hawaii bottomfish revisited: incorporating economics into total allowable catch management. U.S. Department of Commerce, NOAA Technical Memo, NOAA-TM-NMFS-PIFSC-20. 19 pgs+Appendix p.
  76. 76. Martel SJD, Korman J, Darcy M, Christensen LB, Zeller D (2011) Status and trends of the Hawaiian bottomfish stocks: 1948–2006. A report submitted under Contract No. JJ133F-06-SE-2510 September 2006. Pacific Islands Fish. Sci. Cent., Natl. Mar. Fish. Ser., NOAA, Honolulu, HI 96822-2396. Pacific Islands Fish. Sci. Cent. Admin. Rep. H-11-02C.
  77. 77. Bowen BW, Karl SA (2007) Population genetics and phylogeography of sea turtles. Molecular Ecology 16: 4886–4907.
  78. 78. Waples RS, Hendry AP (2008) Evolutionary perspectives on salmonid conservation and management. Evolutionary Applications 1: 183–188.
  79. 79. Selkoe KA, Watson JR, White C, Ben Horin T, Iacchei M, et al. (2010) Taking the chaos out of genetic patchiness: seascape genetics reveals ecological and oceanographic drivers of genetic patterns in three temperate reef species. Molecular Ecology 19: 3708–3726.
  80. 80. White C, Selkoe KA, Watson J, Siegel DA, Zacherl DC, et al. (2010) Ocean currents help explain population genetic structure. Proceedings of the Royal Society B 277: 1685–1694.