Coral Reefs (2012) 31:839–851 DOI 10.1007/s00338-012-0907-y

REPORT

Twisted sister species of pygmy angelfishes: discordance between taxonomy, coloration, and phylogenetics Joseph D. DiBattista • Ellen Waldrop • Brian W. Bowen • Jennifer K. Schultz • Michelle R. Gaither • Richard L. Pyle Luiz A. Rocha

•

Received: 19 November 2011 / Accepted: 9 April 2012 / Published online: 5 May 2012 Ó Springer-Verlag 2012

Abstract The delineation of reef fish species by coloration is problematic, particularly for the pygmy angelfishes (genus Centropyge), whose vivid colors are sometimes the only characters available for taxonomic classification. The Lemonpeel Angelfish (Centropyge flavissima) has Pacific and Indian Ocean forms separated by approximately 3,000 km and slight differences in coloration. These disjunct populations hybridize with Eibl’s Angelfish (Centropyge eibli) in the eastern Indian Ocean and the Pearl-Scaled Angelfish (Centropyge vrolikii) in the western Pacific. To resolve the evolutionary history of these species and color morphs, we employed mitochondrial DNA (mtDNA) cytochrome b and three nuclear introns (TMO, RAG2, and S7). Phylogenetic analyses reveal three deep mtDNA lineages (d = 7.0–8.3 %) that conform not to species designation or color morph but to geographic region: (1) most Pacific C. flavissima plus C. vrolikii, (2) C. flavissima from the Society Islands in French Polynesia, and (3) Indian Ocean C. flavissima plus C. eibli. In contrast, the nuclear introns each show a cluster of closely related alleles, with frequency differences between the three geographic groups. Hence, the mtDNA phylogeny Communicated by Biology Editor Prof. Philip Munday J. D. DiBattista (&)  E. Waldrop  B. W. Bowen  J. K. Schultz Hawai’i Institute of Marine Biology, P.O. Box 1346, Kane’ohe, HI 96744, USA e-mail: [email protected] M. R. Gaither  L. A. Rocha Section of Ichthyology, California Academy of Sciences, 55 Music Concourse Dr, San Francisco, CA 94118, USA R. L. Pyle Bernice P. Bishop Museum, 1525 Bernice St., Honolulu, HI 96817, USA

reveals a period of isolation (ca. 3.5–4.2 million years) typical of congeneric species, whereas the within-lineage mtDNA UST values and the nuclear DNA data reveal recent or ongoing gene flow between species. We conclude that an ancient divergence of C. flavissima, recorded in the non-recombining mtDNA, was subsequently swamped by introgression and hybridization in two of the three regions, with only the Society Islands retaining the original C. flavissima haplotypes among our sample locations. Alternatively, the yellow color pattern of C. flavissima may have appeared independently in the central Pacific Ocean and eastern Indian Ocean. Regardless of how the pattern arose, C. flavissima seems to be retaining species identity where it interbreeds with C. vrolikii and C. eibli, and sexual or natural selection may help to maintain color differences despite apparent gene flow. Keywords Centropyge  Color variation  Coral reef fish  Hybridization  Incomplete lineage sorting  Mitochondrial DNA  Nuclear introns

Introduction The delineation of species is important for preserving biodiversity, and yet, this process is fraught with difficulties (Coyne and Orr 2004). The problem is particularly acute in reef fishes, whose conspicuous colors are often the sole character available to distinguish taxa. Coloration in fishes serves various functions, including social communication, camouflage, and mimicry, but it can also indicate reproductive isolation if it serves as a cue in mate recognition (Randall 1998; McMillan et al. 1999). If sexual selection is strong (Puebla et al. 2007) or ecological differentiation is rapid (Ramon et al. 2003; Choat 2006; Rocha and Bowen 2008), color polymorphisms in incipient reef fish species

123

840

may accumulate faster than genetic differences at neutral markers, including mitochondrial DNA (mtDNA) genes (Bowen et al. 2006). Therefore, as a taxonomic tool, color is much more reliable when coupled with biogeographic, genetic, or additional morphological data (Rocha 2004). To explore the relationship between coloration, taxonomy, and phylogeographic patterns, we surveyed the Lemonpeel Angelfish (Centropyge flavissima [Cuvier 1831]) throughout its Indo-Pacific range using mtDNA cytochrome b (cyt b) and three nuclear introns (TMO, RAG2, and S7). C. flavissima inhabits shallow lagoons or seaward reefs (Allen et al. 1998) and has a long pelagic larval duration (PLD) relative to other genera of angelfish (PLD = 28–32 days; Thresher and Brothers 1985). C. flavissima is found at Christmas (S10°300 , E105°400 ) and Cocos-Keeling (S12°100 , E96°520 ) Islands in the Indian Ocean and in the Western and Central Pacific Ocean (east to the Marquesas Islands and north to the Line Islands but not Hawaii; see Fig. 1), but despite its high dispersal potential, it is not found in most of the species rich Coral Triangle region between the extremes of its range. This disjunct distribution is coupled with subtle differences in coloration; the blue ring around the eye characteristic of Pacific C. flavissima is notably absent from Indian Ocean populations, and a distinct black bar on the posterior margin of the operculum is only present in the Pacific fish (Fig. 2; Allen et al. 1998). A blue-rimmed black ocellate spot on the side of the body during the juvenile stage of this species also appears to persist in larger juveniles of the Indian Ocean form, compared with the Pacific Ocean form (see photo on p. 131 of Allen and Steene 1987). Our first aim is therefore to determine whether these color differences represent separate populations or deeper evolutionary lineages. Evidence from a molecular phylogeny of angelfishes (Bellwood et al. 2004) indicates that the genus Centropyge is not monophyletic. However, apart from coloration, C. flavissima is indistinguishable from two other species within this genus, Eibl’s Angelfish (Centropyge eibli Klausewitz 1963) and the Pearl-Scaled Angelfish (Centropyge vrolikii [Bleeker 1853]). Although previous research focused on the number of lateral-line or vertical scales as diagnostic characters, these often vary with size of the specimen and are therefore unreliable (Pyle 2003). C. eibli is found exclusively in the eastern Indian Ocean (including parts of Indonesia) and Northwestern Australia, whereas C. vrolikii is distributed throughout the tropical western Pacific, from the Coral Triangle east to the Marshall Islands and south to Vanuatu and the Great Barrier Reef (Fig. 1). C. flavissima regularly hybridizes with these two species in regions of overlap (Pyle and Randall 1994; Hobbs et al. 2009). The propensity of species in this genus to hybridize in areas of sympatry, along with their outlined distributions and morphological similarity, led Pyle (1992) to suggest that they may form a single monophyletic group of recently

123

Coral Reefs (2012) 31:839–851

OKI PAL PAU

CAR PHO

INDI COC

XMA

GBR FIJ

KIR TOK NUK MOR

Species distribution

= C. eibli = C. flavissima = C. vrolikii

Fig. 1 Scaled map indicating collection sites (site abbreviations are provided in Table 1) for Centropyge flavissima (open circles) in the Indo-Pacific. Additional specimens from two related angelfish species, Centropyge eibli (filled black circles) and Centropyge vrolikii (filled gray circles), which have been shown to hybridize in areas of overlap with C. flavissima, were collected opportunistically (additional site abbreviations include the Caroline Islands [CAR], the Great Barrier Reef, Australia [GBR], Indonesia [IND], Okinawa, Japan [OKI], and the Republic of Palau [PAU])

diverged species. In addition, a molecular phylogeny of the Centropyge genus (M.R. Gaither et al. pers. comm.) unambiguously supports sister relationships (monophyly) among these three species. Therefore, our second goal is to test whether these three Indo-Pacific angelfish species represent old lineages that have secondary contact or whether they form a cluster of species that diverged recently.

Materials and methods Sample collection A total of 271 C. flavissima were collected at nine sites while scuba diving or snorkeling between 2005 and 2010 (Fig. 1). Specimens of C. eibli (N = 7) and C. vrolikii (N = 14) were also collected from several locations (Fig. 1). Tissue samples were preserved in saturated saltDMSO (Seutin et al. 1991). Total genomic DNA was extracted using the ‘‘HotSHOT’’ protocol (Meeker et al. 2007) and subsequently stored at -20 °C. Mitochondrial DNA analysis A 594-base pair (bp) segment of the mtDNA cyt b gene was amplified using modified primers, which were

Coral Reefs (2012) 31:839–851

841

Fig. 2 ‘‘Pure’’ Centropyge flavissima (a Pacific Ocean morph, Marshall Islands; d Indian Ocean morph, Christmas Island, Australia), Centropyge vrolikii (c Indonesia), and Centropyge eibli (f Indonesia), in addition to some hybrids between these species (b C. flavissima 9 C.

vrolikii, Guam; e C. vrolikii 9 C. eibli, Indonesia). Note the characteristic blue ring around the orbit of the Pacific Ocean C. flavissima specimen (a), which is absent from the Indian Ocean C. flavissima specimen (d). Photo credit: Luiz Rocha

designed for these species (CFLM_FOR: 50 -TCCCTCC AACATTTCAGCAT-30 ; CFLM_REV: 50 -TCTGGATCTC CAAGCAGGTT-30 ). Polymerase chain reaction (PCR) mixes contained 7.6 ll of BioMix solution (BioMix Red; Bioline Ltd., London, UK), 0.26 lM of each primer, and 5–50 ng template DNA in 15 ll total volume. PCRs also used an initial denaturing step at 95 °C for 3 min, then 35 cycles of amplification (30 s of denaturing at 94 °C, 45 s of annealing at 62 °C, and 45 s of extension at 72 °C), followed by a final extension at 72 °C for 10 min. PCR products were visualized through 1.5 % agarose gel electrophoresis and purified by incubating with exonuclease I and shrimp alkaline phosphatase (ExoSAP; USB, Cleveland, OH, USA) at 37 °C for 60 min, followed by 85 °C for 15 min. All samples were sequenced in the forward direction (and reverse direction for questionable cyt b haplotypes [N = 6]) with fluorescently labeled dye terminators (BigDye version 3.1, Applied Biosystems Inc., Foster City, CA, USA) and analyzed using an ABI 3130XL Genetic Analyzer (Applied Biosystems). The sequences were aligned, edited, and trimmed to a uniform length using Geneious Pro vers. 4.8.4 (Drummond et al. 2009); unique mtDNA cyt b haplotypes were deposited in GenBank (accession numbers: JQ914310–JQ914394). jModelTest vers. 1.0.1 (Posada 2008; but also see Guindon and Gascuel 2003) was used with an Akaike information criterion (AIC) test to determine the best nucleotide substitution model for our dataset; the

TIM1 ? G (Posada 2003) model with a gamma parameter of 0.12 was here selected. To evaluate phylogenetic relationships among cyt b haplotypes, we constructed neighbor-joining (NJ) and maximum-likelihood (ML) trees using PAUP* vers. 4.0 (Swofford 2000). For comparison, we constructed a Bayesian tree using MRBAYES (Ronquist and Huelsenbeck 2003) implemented in Geneious Pro. Bootstrap support values for NJ and ML trees were calculated using default settings with 10,000 replicates; only nodes with bootstrap values [50 % were considered. The Bayesian MCMC search strategy consisted of four heated, 1 million step chains with an initial burn-in of 100,000 steps. A single Centropyge bicolor sample (Genbank accession number: JQ914309) was used to root the tree. We calculated divergence between mitochondrial lineages (here denoted d) as the uncorrected pairwise sequence distance between lineages minus the pairwise sequence distance within a lineage using ARLEQUIN vers. 3.1 (Excoffier et al. 2005). The evolutionary relationship among all angelfish haplotypes was further explored with an unrooted network constructed with NETWORK vers. 4.5.1.0 (www.fluxusengineering.com/network_terms.htm) using a median-joining algorithm and default settings (as per Bandelt et al. 1999). ARLEQUIN was used to calculate haplotype (h) and nucleotide diversity (p), as well as to test for range-wide patterns of population structure for C. flavissima. Initially, global UST, which incorporates sequence divergence in

123

842

Coral Reefs (2012) 31:839–851

addition to allele frequency differentiation, was estimated using analysis of molecular variance (AMOVA; Excoffier et al. 1992); deviations from null distributions were tested with nonparametric permutation procedures (N = 99,999). Subsequently, pairwise UST statistics were generated to identify particular sites associated with genetic partitioning. We controlled for false discovery rate using the method of Narum (2006). Deviations from neutral sequence evolution were assessed with Fu’s FS (Fu 1997) for each population using ARLEQUIN; significance was tested with 99,999 permutations. Each site (and lineage) was also fitted with the population parameter s in order to estimate the time since the most recent population (or lineage) expansion (as per Rogers and Harpending 1992). Time since expansion was estimated using the equation s = 2lt, where t is the age of the population (or lineage) in generations, and l is the mutation rate per generation for the sequence (l = number of bp 9 divergence rate within a lineage 9 generation time in years). We used a range of cyt b mutation rates, available from previous fish studies: 1 % per million years (MY) within lineages (Bowen et al. 2001; Reece et al. 2010) or 1.55 % per MY within lineages (Lessios 2008). While generation time is unknown for our study species, we conditionally used the equation T = (a ? x)/2, where a is the age at first reproduction, and x is the age at last reproduction (or longevity; Pianka 1978). We therefore obtained a generation time of 4 years for C. flavissima based on the study of larval stages in captivity for this species (20 months to reproductive maturity and 7-year longevity; Frank Baensch, Reef Culture Technologies, LLC, Oahu, HI). Although absolute expansion times should be interpreted with caution owing to mutation rate and generation time estimates, relative comparisons are robust to such approximations.

TGGCCGTC-30 ) and S7RPEX2R (50 -AACTCGTCTGGCT TTTCGCC-30 ; Chow and Hazama 1998). Sequences for each nuclear locus were aligned and edited using Geneious Pro. In all cases, the alignment was unambiguous with no frameshift mutations or indels; unique genotypes were deposited in GenBank (accession numbers: JQ914560–JQ914585 [TMO], JQ914395–JQ914456 [RAG2], JQ914457–JQ914559 [S7]). Allelic states of nuclear sequences trimmed to a uniform length, with more than one heterozygous site, were estimated using the Bayesian program PHASE vers. 2.1 (Stephens and Donnelly 2003) as implemented in the software DnaSP vers. 5.0 (Librado and Rozas 2009). We conducted 3 runs in PHASE for each dataset with a burn-in of 10,000, and 100,000 or 200,000 iterations. All runs returned consistent allele identities, and PHASE was able to differentiate most alleles with [85 % probability except at single nucleotide positions in 3 individuals at the TMO locus, 19 individuals at the RAG2 locus, and 28 individuals at the S7 locus (or 3, 17, and 26 % of the samples, respectively), which were excluded from further analysis. Unrooted median-joining networks were produced for each nuclear dataset as outlined above. Although the S7 network was further simplified by removing all singleton alleles (N = 4) to minimize circularity between closely related alleles, this did not influence our overall interpretation. ARLEQUIN was used to test for genetic structure between lineages, color morphs, and previously recognized species. Because jModelTest did not converge on a model of sequence evolution for any nuclear locus, we calculated global and pairwise FST values based on conventional allele frequencies only.

Nuclear gene analysis

Mitochondrial DNA analysis

To provide independent estimates of phylogenetic relationships and ensure genealogical concordance across multiple loci, we sequenced a subset of the specimens at three nuclear genes. Individuals sequenced included specimens from all species and across the three observed mtDNA lineages (Lineage 1, N = 50; Lineage 2, N = 24; Lineage 3, N = 34). Approximately 254 bp of the TMO 4C4 gene was amplified using the modified primers TMO F1 (50 -ACCTCTCATTAAGAAAMGAGTGTTTG-30 ) and TMO R2 (50 -TGCTTCTCAAATTCTTTMACCTS-30 ), 122 bp of the recombination-activating gene 2 (RAG2) was amplified using the modified primers RAG2 2F (50 -SACCTTGTGCTGCAAAGAGA-30 ) and RAG2 2R (50 -GG ATCCCCTTBTCATCCAGA-30 ), and 120 bp of the first intron of the S7 ribosomal protein (S7) gene was amplified using the primers S7RPEX1F (50 -TGGCCTCTTCCT

Phylogenetic analysis revealed that samples partitioned into three reciprocally monophyletic lineages, based on cyt b sequence data: (1) C. flavissima and C. vrolikii sampled at most Pacific Ocean sites, which shared six haplotypes between the species, (2) C. flavissima sampled at Moorea, and (3) C. flavissima and C. eibli sampled at Indian Ocean sites, which shared the most common haplotype (Fig. 3). Lineage 1 was 7.2 and 7.0 % divergent from Lineage 2 and Lineage 3, respectively (which represents ca. 3.6 or 3.5 MY of separation, based on 2 % per MY between lineages), whereas Lineage 2 was 8.3 % divergent from Lineage 3 (ca. 4.2 MY of separation). Sequence divergence within each lineage, on the other hand, was low (Lineage 1, d = 0.17–0.26 %; Lineage 2, d = 0.20–0.31 %; Lineage 3, d = 0.15–0.22 %). The outgroup, C. bicolor, was 12.0–14.0 % divergent from the C. flavissima complex.

123

Results

Coral Reefs (2012) 31:839–851

Median-joining haplotype networks were consistent with a scenario of minimal genetic differentiation within each lineage and no shared haplotypes between geographic groupings (Fig. 4). The only exception was Nuku Hiva in the Marquesas Archipelago (within Lineage 1), which did not share a haplotype with any other Pacific location and was distinguished by two diagnostic mutations (at bp 300 and bp 567). C. flavissima cyt b sequence data revealed 43, 19, and 17 haplotypes for Lineages 1, 2, and 3, respectively, with h = 0.22 to 0.94 and p = 0.0004 to 0.0043 for individual sample locations (Table 1). Nucleotide diversity was almost twice as high in Lineage 2 (Moorea) versus Lineage 3 (Indian Ocean), which cannot be explained by a greater sampling effort in the former (N = 45 vs. N = 66). Tests of neutrality revealed negative and significant Fu’s FS values at all sample sites (Fu’s FS = -18.28 to -2.58; Table 1). Our estimates of s yielded time of expansion approximations for each lineage, with Lineage 3 (63,000–97,600 years) expanding more recently than either Lineage 1 (100,400–155,600 years) or Lineage 2 (131,700–204,200 years), although there was variability among individual sites within lineages (Table 1). AMOVA confirmed the geographic grouping based on mtDNA lineages (overall UST = 0.98, P \ 0.001; Table 2), with 97 % of the variation in haplotype diversity explained by these three groups. Population pairwise tests also revealed that mtDNA haplotypes were significantly different in 25 of 36 comparisons (all P \ 0.001; Table 3), with estimates of UST ranging from -0.01 to 0.99. All UST values were significant between geographic regions, but Nuku Hiva (Marquesas Islands) was the only site that was significantly different from other sites within regions (UST = 0.70 to 0.81, all P \ 0.001; Table 3). Nuclear gene analysis Based on nuclear sequences from 108 angelfish specimens, 15 variable sites yielded 18 alleles at the TMO locus, 8 variable sites yielded 13 alleles at the RAG2 locus, and 7 variable sites yielded 12 alleles at the S7 locus. Medianjoining networks for all three nuclear genes revealed common alleles at each locus that were shared among every sampling site or putative species (Fig. 5). We did, however, detect significant shifts in allele frequencies among samples from Indian Ocean sites, Moorea, and all other Pacific Ocean sites (TMO, FST = 0.062, P \ 0.001; RAG2, FST = 0.024, P = 0.028; S7, FST = 0.062, P = 0.0013). We found few allele frequency differences, however, among recognized species (TMO, FST = 0.26, P \ 0.001; RAG2, FST = 0.040, P = 0.084; S7, FST = 0.025, P = 0.12), and none between C. flavissima color morphs (TMO, FST = 0.029, P = 0.33;

843

RAG2, FST = 0.028, P = 0.29; S7, FST = 0.041, P = 0.22), indicating that these markers may be in the process of segregating by location, albeit much more slowly than the mtDNA.

Discussion Our genetic survey of C. flavissima color morphs and two closely related species (C. vrolikii and C. eibli) indicates that previous taxonomic divisions are not compatible with a molecular phylogenetic hypothesis. The mtDNA analyses identified three monophyletic lineages that grouped by sampling location rather than coloration or species designation. C. flavissima and C. vrolikii sampled at Pacific Ocean sites grouped together in Lineage 1, C. flavissima sampled at Moorea grouped in Lineage 2, and C. flavissima and C. eibli sampled at Indian Ocean sites grouped in Lineage 3. Our mtDNA results confirm that the two C. flavissima color morphs, separated by approximately 3,000 km in adjacent ocean basins, are following independent evolutionary trajectories with an approximate divergence time of 3.5 MY. Genetic differentiation at the three nuclear genes indicates concordant separation between these three regional groups, although at the level of allele frequency differences only. This is consistent with recent work in other reef fish systems (e.g., Cleaner Wrasse, Labroides dimidiatus, Drew et al. 2008; Browncheek Blenny, Acanthemblemaria crockery, Lin et al. 2009), where location but not phenotype dictates the genetic affinity of color morphs. Discordance between genetic divergence and coloration is well documented in butterflyfishes (family Chaetodontidae: McMillan and Palumbi 1995), the sister family to marine angelfishes, and in hamlets (Ramon et al. 2003; Garcia-Machado et al. 2004); both groups having bright coloration. Counter examples exist as well, wherein similar coloration masks evolutionary divergence in damselfishes (Pomacentridae: Planes and Doherty 1997a, b; Bernardi et al. 2002; McCafferty et al. 2002; Rocha 2004; Drew et al. 2010; Leray et al. 2010), groupers (Serranidae: McCartney et al. 2003; Craig et al. 2009), and wrasses (Labridae: Rocha et al. 2005). In most cases, however, color morphs do correspond to genetic partitions (Randall and Rocha 2009). Centropyge angelfishes stand out as an evolutionary enigma, even against this backdrop of discordant taxonomy, coloration, and genetics in reef fishes. All phylogeographic studies published on Centropyge to date (including this one) show discordance between color morphs (or coloration-based taxonomy) and genetic partitions. In the Atlantic, color differences separate three described species inhabiting the Caribbean, Brazil, and mid-Atlantic ridge, but these species share mtDNA

123

844

Coral Reefs (2012) 31:839–851

cflm1139 (31) cflm1151 (1) cflm1154 (1) cflm1157 (1) cflm1171 (1) cvr1005 (CAR; 1) + cflm1207 (5) cflm15 (1) cflm1093 (1) cflm1251 (1) cflm1050 (2) cflm1257 (1) cvr4 (CAR/GBR/OKI; 8) + cflm1206 (54) cvr3 (GBR; 1) + cflm1047 (1) cvr1006 (CAR; 1) + cflm1058 (1) cvr7 (OKI; 1) cvr1001 (PAU; 1) + cflm1205 (6) cflm1212 (1) cflm17 (2) cflm1095 (1) cflm1097 (1) cflm1100 (4) cflm1101 (2) cflm1102 (1) cflm1053 (3) cflm1063 (1) cflm1066 (1) cflm1220 (1) cflm21 (2) cflm1254 (1) cflm1258 (1) cflm1203 (4) cflm1216 (1) cvr1002 (CAR; 1) + cflm1103 (4) cflm1211 (6) cflm1214 (3)

0.4 substitutions/site

100/100/1.00

Lineage 1 (Pacific Ocean & C. vrolikii)

cflm1215 (1) cflm1217 (3) cflm1045 (1) cflm1065 (1) cflm1090 (1) cflm1076 (1) cflm1272 (1) cflm1040 (2) cflm1221 (1) cflm1225 (10)

73/67/0.91

100/100/1.00

cflm1226 (1) cflm1237 (1) cflm1239 (19) cflm1248 (1) cflm1179 (1) cflm9 (1) cei1002 (XMA; 1) cei2 (IND; 1) cflm1230 (3) cflm12 (1) cflm1224 (19) + cei1 (IND/XMA; 2) cflm1227 (1) cflm1234 (2) cflm1236 (1) cflm1240 (1) cflm1188 (1) cflm1194 (2) cflm1200 (1) cei1001 (XMA; 1)

Lineage 3 (Indian Ocean and C. eibli)

cei1003 (XMA; 1) cei3 (IND; 1)

100/100/1.00

cflm1105 (1) cflm1123 (5) cflm1 (1) cflm1104 (17) cflm1110 (1) cflm1111 (3) cflm1112 (1) cflm1113 (1) cflm1116 (2) cflm1117 (1) cflm1121 (1) cflm1124 (2) cflm1131 (2) cflm1133 (1) cflm1135 (1) cflm1138 (1) cflm13 (2) cflm26 (1) cflm31 (1)

Lineage 2 (Moorea, French Polynesia)

Centropyge bicolor

haplotypes, and the Brazilian and Caribbean species are not distinguishable even at the population level (Bowen et al. 2006). The Flame Angelfish (Centropyge loricula) in the

123

central Pacific maintains distinct color morphs in different parts of its range yet shows high gene flow between these regions (Schultz et al. 2007). These observations may be

Coral Reefs (2012) 31:839–851

845

bFig. 3 Phylogenetic relationship among mitochondrial DNA cyto-

chrome b haplotypes (594 base pairs) for Centropyge flavissima specimens (cflm; N = 271) collected in the Indo-Pacific based on neighbor-joining (NJ), maximum-likelihood (ML), and Bayesian (BA) inference. Additional sequences from two closely related angelfish, Centropyge eibli (cei: Christmas Island, Australia [XMA], N = 4; Indonesia [IND], N = 3) and Centropyge vrolikii (cvr: Caroline Islands [CAR], N = 6; Great Barrier Reef, Australia [GBR], N = 4; Okinawa, Japan [OKI], N = 3; Republic of Palau [PAU], N = 1) were also included. Branch support values ([50 %, based on 10,000 replicates) for NJ and ML analysis, and posterior probabilities for BA analysis are shown above the nodes (NJ/ML/ BA). All analyses resulted in identical lineages, and so the BA topology is presented here. Branch lengths are according to indicated scale but the branch leading to the outgroup species, Centropyge bicolor, was here reduced by 50 %. Values in parentheses represent the number of samples for each haplotype, and colors denote collection location as indicated by the embedded key. Each haplotype in the tree was therefore assigned a corresponding rectangle to the right of the figure, and each color denoting a location within that rectangle was proportional to its total frequency for that haplotype

the key to understanding the triad of putative species examined here, as apparently members of this genus tend to preserve color differences despite gene flow. Fig. 4 Median-joining networks showing relationships among mitochondrial DNA, non-singleton cytochrome b haplotypes for Centropyge flavissima specimens (N = 271) collected in the Indo-Pacific. Additional sequences from two closely related angelfishes, Centropyge eibli (cei: Christmas Island, Australia [XMA], N = 4; Indonesia [IND], N = 3) and Centropyge vrolikii (cvr: Caroline Islands [CAR], N = 6; Great Barrier Reef, Australia [GBR], N = 4; Okinawa, Japan [OKI], N = 3; Republic of Palau [PAU], N = 1), were also included. Each circle represents a haplotype and its size is proportional to its total frequency. Branches or black crossbars represent a single nucleotide change, open circles represent unsampled haplotypes, and colors denote collection location as indicated by the embedded key. The network was separated into the three distinct lineages (a Lineage 1; b Lineage 2; c Lineage 3)

a

What conditions can explain the maintenance of color differences in the face of gene flow? The most likely explanation is that natural or sexual selection is acting to conserve these color differences. For many reef fishes, coloration is a key character involved in mate recognition and therefore subject to strong sexual selection (Seehausen et al. 1997, 1999). As one example, strong preference for mating with their own morphotype (assortative mating) appears to have played a role in maintaining evolutionarily stable color polymorphisms in Caribbean hamlets (Hypoplectrus complex; Puebla et al. 2008; Holt et al. 2011). Selective pressures that are not directly related to mate choice (i.e., predation, habitat preference, or territoriality) can also reinforce differences in color patterns among butterflyfishes in the face of ongoing gene flow (McMillan et al. 1999). Indeed, predation against novel color types is an important factor in reducing hybrid success in other vertebrate systems (Langham 2007) and may factor into the maintenance of regional color morphs here. Although our three study species inhabit similar depths and habitats (Hobbs et al. 2010), C. flavissima is found almost

Cyt b - Lineage 1 (Pacific Ocean & C. vrolikii)

Moorea Nuku Hiva X-mas Island, Pacific Ocean Palmyra Tokelau Islands Phoenix Islands Fiji X-mas Island, Indian Ocean Cocos-Keeling

b

Cyt b - Lineage 2 (Moorea, French Polynesia)

Centropyge eibli (cei) Centropyge vrolikii (cvr)

c

Cyt b - Lineage 3 (Indian Ocean & C. eibli )

123

846

Coral Reefs (2012) 31:839–851

Table 1 Molecular diversity indices for Centropyge flavissima based on mitochondrial DNA (cytochrome b) sequence data for all sampling locations, including the three identified Indo-Pacific lineages (also see Fig. 3) Collection locality

N

HN

HU

Time since expansion (Years)

Haplotype diversity (h ± SD)

Nucleotide diversity (p ± SD)

Fu’s FS

Pacific Ocean Christmas Island, Line Islands (KIR)

43

19

8

69,700–108,000

0.78 ± 0.066

0.0026 ± 0.0018

218.28a

Fiji (FIJ) Nuku Hiva, Marquesas Islands (NUK)

19 35

12 5

4 5

106,500–165,100 126,300–195,700

0.94 ± 0.035 0.22 ± 0.092

0.0043 ± 0.0027 0.0004 ± 0.0005

26.97 24.74

Palmyra Atoll, Line Islands (PAL)

30

11

5

73,700–114,200

0.67 ± 0.093

0.0023 ± 0.0016

26.96

Phoenix Islands (PHO)

24

15

6

89,400–138,600

0.87 ± 0.067

0.0035 ± 0.0023

212.26

Tokelau Islands (TOK)

9

6

1

82,900–128,400

0.83 ± 0.130

0.0030 ± 0.0022

22.58

All samples—Lineage 1

160

43

43

100,400–155,600

0.85 ± 0.023

0.0035 ± 0.0022

227.30

45

19

19

131,700–204,200

0.85 ± 0.049

0.0043 ± 0.0026

211.93 26.32

Moorea, French Polynesia (MOR) All samples—Lineage 2 Indian Ocean Christmas Island, Aus. (XMA)

27

11

6

70,100–108,600

0.85 ± 0.042

0.0027 ± 0.0018

Cocos-Keeling Islands, Aus. (COC)

39

11

6

56,600–87,700

0.78 ± 0.047

0.0027 ± 0.0018

-5.01

All samples—Lineage 3

66

17

17

63,000–97,600

0.82 ± 0.029

0.0026 ± 0.0017

211.13

Time since the most recent population expansion was calculated using a range of mutation rates (1–1.55 % per million years within lineages; Bowen et al. 2001; Lessios 2008; Reece et al. 2010) and a generation time of four years (see ‘‘Methods’’) N sample size, HN number of haplotypes, HU number of unique haplotypes a

Numbers in bold are significant, P \ 0.02 (see Fu 1997)

Table 2 Results of the analysis of molecular variance (AMOVA) based on mitochondrial DNA cytochrome b sequence data for Centropyge flavissima (N = 271) SS

Variance components

% Variation

UCT USC

P-value

UST

P-value

2

7147.24

46.51

97.43

0.97

0.004

0.98

\0.001

0.32

\0.001

6 262

68.79 217.36

0.40 0.83

0.83 1.74

Source

df

Among groups Among populations (within groups) Within populations

Groups were based on the three Indo-Pacific lineages identified by phylogenetic reconstruction (see Fig. 3). UCT is the group variance component relative to total variance, USC is the between sample within group variance component divided by the sum of itself and within sample variance, and UST is the sum of the variance due to group and sample within group divided by the total variance df degrees of freedom, SS sum of squares

exclusively on reefs around low oceanic islands, whereas C. vrolikii and C. eibli predominate at high islands or continental shelves, indicating that there may be some niche partitioning in areas of overlap. Our findings include deep (reciprocally monophyletic) differentiation in mtDNA, coupled with weak populationlevel differentiation at nuclear loci, which can certainly be explained by the non-recombining nature of mtDNA inheritance and the slower mutation rate at introns relative to mtDNA. Two additional (not mutually exclusive) factors require consideration here: (1) incomplete lineage sorting and (2) hybridization and introgression. These two explanations are difficult to tease apart but we discuss each possibility. It should be noted that natural selection for

123

particular mtDNA types can also produce such a pattern, although such selective sweeps are thought to be rare (Karl et al. 2012). Incomplete lineage sorting occurs during the transitional stage when evolutionary lineages begin to diverge (Avise 2004). The haploid inheritance of mtDNA yields a fourfold lower effective population size (Ne) relative to nuclear DNA, such that isolated mtDNA lineages are expected to drift to reciprocal monophyly in Ne generations on average, whereas diploid nuclear loci will attain this evolutionary divergence in 4Ne generations on average (Tavare´ 1984). A divergence time of 4 MY (based on our mtDNA cyt b molecular clock) seems sufficient for these lineages to completely sort at both mtDNA and nuclear genes, although no comparable

0.98 (\0.001) Site abbreviations are outlined in Table 1 Numbers in bold are significant, P B 0.012 (as per Narum 2006)

0.99 (\0.001) NUK

b

0.98 (\0.001) MOR

a

– 0.98 (\0.001)

–

0.70 (\0.001) 0.75 (\0.001) 0.81 (\0.001) 0.70 (\0.001) 0.71 (\0.001)

0.98 (\0.001)

-0.01 (0.75) -0.02 (0.68)

0.97 (\0.001) 0.97 (\0.001)

-0.02 (0.96) 0.98 (\0.001) KIR

0.98 (\0.001)

0.98 (\0.001) PAL

0.98 (\0.001)

0.98 (\0.001) TOK

0.04 (0.026)

– –

-0.03 (0.81) -0.01 (0.73) 0.04 (0.040)

0.98 (\0.001) PHO

0.98 (\0.001)

-0.03 (0.94)

– 0.01 (0.25)

-0.03 (0.70)

0.98b (\0.001) FIJ

0.98 (\0.001)

– 0.05 (0.032) XMA

0.98 (\0.001)

–

0.98 (\0.001)

– COC

0.97 (\0.001)

KIR PAL TOK PHO FIJ XMA COC

0.99 (\0.001)

MOR

–

NUK

847

Locationa

Table 3 Matrix of population pairwise UST values (below diagonal) with associated P-values in parentheses, based on mitochondrial DNA cytochrome b sequence data for Centropyge flavissima (N = 271) sampled at sites throughout the Indo-Pacific

Coral Reefs (2012) 31:839–851

molecular clock is available for the latter. That said, the introns in this study show weak population structure between regions and extensive sharing of alleles, indicating that nuclear lineage sorting is not proceeding toward fixation of alternate states. While incomplete lineage sorting almost certainly contributes to our findings, it is not sufficient to explain the discordance between mtDNA and nuclear introns. The alternate explanation, hybridization and introgression, can explain the observed discordance between mtDNA and nuclear DNA. Because pygmy angelfish are primarily distinguished on the basis of coloration and are also highly prized in the aquarium trade, hybrids between these species tend to be both noticed and documented. Hence, these hybrids are among the best characterized for any tropical reef fish family (Pyle and Randall 1994; but also see Yaakub et al. 2006; see Fig. 2). C. flavissima 9 C. vrolikii hybrids (documented genetically by L.A. Rocha) are regularly exported through the aquarium trade from the Marshall Islands, Pohnpei, Guam, Kosrae, the Ryukyu Islands, and Vanuatu (Takeshita 1976) and have been reported from almost all locations where these two species co-occur. At these overlapping sites, individual fish range from nearly ‘‘pure’’ C. flavissima to nearly ‘‘pure’’ C. vrolikii. No hybrids have been reported from New Caledonia, the Solomon Islands, and Palau (where only C. vrolikii is present), or the Society Islands, Line Islands, Phoenix Islands, and Marquesas Islands (where only C. flavissima is present). Although hybrids have not been reported (to our knowledge) from areas of sympatry in the Ogasawara Islands and the Great Barrier Reef, this could be attributed to insufficient sampling effort or very low abundance of one or the other species (Pyle and Randall 1994; D. Bellwood pers. comm.). The two species that hybridize with C. flavissima, C. eibli, and C. vrolikii, also hybridize with each other in Indonesia (Pyle and Randall 1994), and C. flavissima 9 C. eibli hybrids are common at Christmas and CocosKeeling Islands (400 and 1,000 km southwest of Indonesia; Hobbs et al. 2009), which is at the boundary of the Western Indo-Pacific and the Central Indo-Pacific ecoregions (Spalding et al. 2007). The sharing of mtDNA haplotypes between recognized species (this study), the spectrum of intermediate color forms (Pyle and Randall 1994), and the presence of areas where hybrids outnumber parental forms (Pyle and Randall 1994) are evidence for hybridization that extends beyond the F1 generation. Moreover, rarity of conspecific partners and a polygynous mating system (i.e., harems with a single male and two to seven or more females; Moyer and Nakazono 1978; Moyer 1990), likely promote hybridization in these fish. We therefore conclude that hybridization has played a major role in shaping the genetic architecture of the three angelfish lineages

123

848 Fig. 5 Median-joining networks showing relationships among nuclear DNA alleles (a TMO, 254 base pairs, N = 145; b RAG2, 122 base pairs, N = 160; c S7, 120 base pairs, N = 138) based on a subset of all angelfish samples collected in this study. Each circle represents an allele and its size is proportional to its total frequency. Branches or black crossbars represent a single nucleotide change, open circles represent unsampled alleles, and colors denote collection location as indicated by the embedded key. All singleton alleles (N = 4) were removed from the S7 analysis in order to minimize circularity between closely related alleles

Coral Reefs (2012) 31:839–851

a TMO

Moorea Nuku Hiva X-mas Island, Pacific Ocean Palmyra Tokelau Islands Phoenix Islands Fiji X-mas Island, Indian Ocean Cocos-Keeling

b RAG2

Centropyge eibli (cei) Centropyge vrolikii (cvr)

c S7

observed in this study, although further research on the fertility and viability of hybrids is needed to confirm this conclusion. We here suggest a scenario where the species complex was divided three to four MY ago into ancestral C. eibli in the Indian Ocean, C. vrolikii in the western Pacific, and C. flavissima in the central South Pacific. C. flavissima subsequently extended its range to the western Pacific, where it hybridized with C. vrolikii, and the eastern Indian Ocean where it hybridized with C. eibli. The end result is three lineages: C. eibli found in the Indian Ocean (whose haplotypes are mixed with Indian Ocean C. flavissima due to introgression), C. vrolikii found everywhere in the western Pacific (whose haplotypes are mixed with western Pacific C. flavissima, also due to introgression), and C. flavissima found in the Society Islands, which may be the last refugium of pure individuals in this lineage. The fact that C. flavissima currently inhabits the Phoenix Islands without C. vrolikii, yet still has C. vrolikii mtDNA, supports a scenario of introgression. Even though precise reasons for this ancient isolation could not be identified,

123

similarly deep mtDNA divergences have been observed in other reef fish: two lineages coalescing between 2.9 and 5.5 MY ago were identified for Naso vlamingii (Klanten et al. 2007), and three lineages coalescing between 2.0 and 5.0 MY ago were observed in Naso brevirostris (Horne et al. 2008). In both cases, the authors invoke ancient periods of isolation likely caused by sea level fluctuations to explain the observed genetic divergences. Indeed, the intensification of glaciation 2.7 MY ago is one of several factors that altered and accelerated surface currents in the Indo-Pacific (Ivanova 2009). We suggest that these processes started the ancient differentiation in our study group, with one key distinction: the Naso lineages show no color or morphological difference and are completely mixed in all localities, whereas in our case, the lineages are still segregated geographically, but they do not match the color (and species) boundaries. A second and perhaps equally plausible (although hard to test) scenario is the possibility that the yellow coloration of C. flavissima originated independently in the Indian and Pacific oceans (Pyle 2003). In this scenario, the ancestral

Coral Reefs (2012) 31:839–851

C. eibli in the Indian Ocean and the ancestral C. vrolikii in the Pacific Ocean each gave rise to a xanthic form. Indeed, the aquarium trade has documented a long list of xanthic fish variants that are cultivated for their beauty and novelty, including a wild-caught xanthic strain of the Dusky Angelfish, Centropyge multispinis (http://www.reefs.org/ forums/topic120708.html). Slight color differences between the Indian and Pacific Ocean ‘‘C. flavissima’’ lend further support to this hypothesis, however, the presence of three genetic lineages instead of two (one C. eibli ? Indian Ocean xanthics and one C. vrolikii ? Pacific Ocean xanthics), and the lack of xanthic C. vrolikii and C. eibli inside their respective ranges, weaken it. All but one of our sampling sites is located in the vast Indo-Polynesian biogeographic province, which stretches from Polynesia to the central Indian Ocean (Briggs and Bowen 2012). Our sample from Nuku Hiva in the Marquesas Archipelago (east of the Society Islands; Fig. 1) represents a distinct biogeographic province with a depauperate fish fauna and high rates of endemism (11.6 % in fishes; Randall 1998). Recent phylogeographic surveys have found genetic differentiation between Society Island and Marquesan populations in several groups of fishes, including snappers (genus Lutjanus; Gaither et al. 2010), surgeonfishes (genus Acanthurus; Planes and Fauvelot 2002), and wrasses (genus Halichoeres; W.B. Ludt et al. pers. comm.). The distinctiveness of Marquesan reef fishes is attributed to a combination of geographic isolation owing to the westerly South Equatorial Current, limited coral reef development, and variable water temperatures due to major upwelling events (Randall 1998; Gaither et al. 2010). Our results reinforce the genetic uniqueness of fishes in the Marquesas Islands, with C. flavissima sampled at Nuku Hiva not sharing haplotypes with any other location (Table 1, Fig. 4). In phylogenetic analyses, however, the Marquesas individuals grouped with the widespread Pacific lineage and not the nearby Society Islands lineage (Fig. 3). A relatively recent colonization event is consistent with the low genetic variation detected at Nuka Hiva despite good sampling effort (N = 35); haplotype or nucleotide diversity was at least three or six times lower here than at any other location. Further sampling is needed at other reef systems in the Society Islands, Tuamotu Islands, and surrounding islands chains (such as Samoa or Tonga) to test our hypotheses of a recent colonization event of the Marquesas Islands by introgressed western Pacific C. flavissima, in addition to the rest of French Polynesia acting as a refugium for the ‘‘pure’’ C. flavissima. Indeed, we identified a single C. flavissima specimen from Tonga (2,000 km west of the Society Islands; GenBank Accession Number: FJ582964.1) that groups with our Moorea samples based on the mtDNA barcoding gene (COI; data not shown), indicating that there are other places in the Central Pacific

849

where C. flavissima and C. vrolikii have not yet fully introgressed. Taxonomic implications Since our phylogeny seems to contradict recognized species boundaries, are these angelfishes valid species? While the recognized angelfish species in the complex studied here might not represent reciprocally monophyletic mtDNA or nuclear intron lineages, they represent stable color forms and the presence of deeply separated mtDNA lineages indicate that those forms have existed for millions of years. The solution proposed by de Queiroz (2007) for dilemmas like the one presented here is simple; use the common element to define the species, and one or more secondary properties as qualifiers to support this designation. In our case, despite the apparent gene flow, these species maintain unique color characteristics and are partitioned into cohesive geographic regions. We therefore suggest that these angelfish remain recognized as taxonomically diagnosable species. In conclusion, the emerging picture of evolution in the pygmy angelfishes (genus Centropyge) includes a number of factors known from other organisms, but which combine here into a unique synthesis of dispersal, hybridization, natural (or sexual) selection, and speciation. First, this group contains members that are good dispersers, with low or no population structure recorded across entire ocean basins (Bowen et al. 2006; Schultz et al. 2007; present study). This is almost certainly due to a pelagic larval stage that readily traverses oceans; Centropyge larvae have been detected in mid-oceanic trawls (MCZ 73476, 73518, 73521, 73531–73532, 73546, 735554–735556, 81683, 82468, 158311, 163525; http://www.mcz.harvard.edu/ Departments/Fish/), notably to the exclusion of many other common and abundant reef fishes (D. Smith and K. Hartel, pers. comm.). Second, this group is known for extensive natural hybridization (Pyle and Randall 1994), a phenomenon that is rare in other groups of reef organisms (Hubbs 1955; Gardner 1997; but see Hobbs et al. 2009). Third, all seven species examined with molecular data retain regional color morphs in the face of gene flow, a signature of natural or sexual selection. At least some species are also sexually dimorphic for color patterns (Allen et al. 1998), adding greater weight to the interpretation that coloration influences mate choice. Fourth, the three species (C. flavissima, C. eibli, and C. vrolikii) would probably be regarded as a single species if coloration was omitted from taxonomic consideration (Pyle 2003). In these circumstances, it is tempting (and defensible) to conclude that these are emerging species, but this is contradicted by the ancient history inscribed in mtDNA, which

123

850

is also concordant with biogeographic partitions observed in other reef fish species (Rocha et al. 2007). The evidence therefore indicates that at one time three species existed but that extensive dispersal and hybridization has rearranged formerly isolated species into semi-isolated color morphs. Like the cichlids of Africa’s rift lakes, the novel aspects of pygmy angelfish evolution will continue to provide insights about the ragged edge of speciation in the oceans. Acknowledgments This research was supported by the National Science Foundation grants OCE-0453167 and OCE-0929031 to BWB, NOAA National Marine Sanctuaries Program MOA No. 2005-008/66882 to R.J. Toonen, and by a Natural Sciences and Engineering Research Council of Canada (NSERC) postgraduate fellowship to JDD. For specimen collections, we thank Kim Andersen, Paul Barber, Larry Basch, David Bellwood, J. Howard Choat, Matthew Craig, Joshua Drew, John Earle, Jeff Eble, Brian Greene, Matthew Iacchei, Stephen Karl, Randall Kosaki, David Pence, and Ross Robertson. We thank Sue Taei at Conservation International, Graham Wragg of the RV Bounty Bay, the Government of Kiribati, including Tukabu Teroroko and the Phoenix Island Protected Area who assisted with Kiribati collections. We also thank Robert Toonen, Serge Planes, Stephen Karl, John Randall, Joann Leong, Patrick Colin, Laura Colin, the Coral Reef Research Foundation, and members of the ToBo lab for their logistic support; we thank the Center for Genomics, Proteomics, and Bioinformatics at the University of Hawaii for their assistance with DNA sequencing. This is contribution no. 1492 from the Hawai’i Institute of Marine Biology and no. 8605 from the School of Ocean and Earth Science and Technology.

References Allen GR, Steene R (1987) Reef fishes of the Indian Ocean. T.F.H. Publications Inc., Neptune City Allen GR, Steene R, Allen M (1998) A guide to angelfishes and butterflyfishes. Odyssey Publishing/Tropical Reef Research, Perth, Australia Avise JC (2004) Molecular markers, natural history, and evolution, 2nd edn. Sinauer Associates, Sunderland, MA Bandelt HJ, Forster P, Ro¨hl A (1999) Median-joining networks for inferring intraspecific phylogenies. Mol Biol Evol 16:37–48 Bellwood DR, Van Hewerden L, Konow N (2004) Evolution and biogeography of marine angelfishes (Pisces: Pomacanthidae). Mol Phylogenet Evol 33:150–155 Bernardi G, Holbrook SJ, Schmitt RJ, Crane NL, DeMartini E (2002) Species boundaries, populations and colour morphs in the coral reef three-spot damselfish (Dascyllus trimaculatus) species complex. Proc R Soc Lond B Biol Sci 269:599–605 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 Bowen BW, Muss A, Rocha LA, Grant WS (2006) Shallow mtDNA coalescence in Atlantic pygmy angelfishes (genus Centropyge) indicates a recent invasion from the Indian Ocean. J Hered 97:1–12 Briggs JC, Bowen BW (2012) A realignment of marine biogeographic provinces with particular reference to fish distributions. J Biogeogr 39:12–30 Choat JH (2006) Phylogeography and reef fishes: bringing ecology back into the argument. J Biogeogr 33:967–968 Chow S, Hazama K (1998) Universal PCR primers for S7 ribosomal protein gene introns in fish. Mol Ecol 7:1255–1256

123

Coral Reefs (2012) 31:839–851 Coyne JA, Orr HA (2004) Speciation. Sinauer Associates, Sunderland, MA Craig MT, Graham RT, Torres RA, Hyde JR, Freitas MO, Ferreira BP, Hostim M, Gerhardinger LC, Andrade AB, Robertson DR (2009) How many species of goliath grouper are there? Cryptic genetic divergence in a threatened marine fish and the resurrection of a geopolitical species. Endanger Species Res 7:167–174 de Queiroz K (2007) Species concepts and species delimitation. Syst Biol 56:879–886 Drew JA, Allen GR, Kaufman L, Barber PH (2008) Endemism and regional color and genetic differences in five putatively cosmopolitan reef fishes. Conserv Biol 22:965–975 Drew JA, Allen GR, Erdmann MV (2010) Congruence between mitochondrial genes and color morphs in a coral reef fish: population variability in the Indo-Pacific damselfish Chrysiptera rex (Snyder, 1909). Coral Reefs 29:439–444 Drummond AJ, Ashton B, Cheung M, Heled J, Kearse M, Moir R, Stones-Havas S, Thierer T, Wilson A (2009) Geneious v4.8, Available from http://www.geneious.com Excoffier R, Smouse P, Quattro J (1992) Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics 131:479–491 Excoffier R, Laval LG, Schneider S (2005) Arlequin ver. 3.0: An integrated software package for population genetics data analysis. Evol Bioinform Online 1:47–50 Fu YX (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147:915–925 Gaither MR, Toonen RJ, Robertson DR, Planes S, Bowen BW (2010) Genetic evaluation of marine biogeographic barriers: perspectives from two widespread Indo-Pacific snappers (Lutjanus spp.). J Biogeogr 37:133–147 Garcia-Machado E, Chevalier Monteagudo PP, Solignac M (2004) Lack of mtDNA differentiation among hamlets (Hypoplectrus, Serranidae). Mar Biol 144:147–152 Gardner JPA (1997) Hybridization in the sea. Adv Mar Biol 31:1–78 Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52:696–704 Hobbs JPA, Frisch AJ, Allen GR, Van Herwerden L (2009) Marine hybrid hotspot at Indo-Pacific biogeographic border. Biol Lett 5:258–261 Hobbs JPA, Jones GP, Munday PL (2010) Rarity and extinction risk in coral reef angelfishes on isolated islands: interrelationships among abundance, geographic range size and specialization. Coral Reefs 29:1–11 Holt BG, Coˆte´ IM, Emerson BC (2011) Searching for speciation genes: Molecular evidence for selection associated with colour morphotypes in the Caribbean reef fish genus Hypoplectrus. PloS ONE [doi:10.1371/journal.pone.0020394.g001] Horne JB, van Herwerden L, Choat JH, Robertson DR (2008) High population connectivity across the Indo-Pacific: Congruent lack of phylogeographic structure in three reef fish congeners. Mol Phylogenet Evol 49:629–638 Hubbs CL (1955) Hybridization between fish species in nature. Syst Zool 4:1–20 Ivanova EV (2009) The global thermohaline circulation. Springer, New York Karl SA, Toonen RJ, Grant WS, Bowen BW (2012) Common misconceptions in molecular ecology: Echos of the modern synthesis. Mol Ecol (in press) Klanten OS, Choat JH, van Hewerden L (2007) Extreme genetic diversity and temporal rather than spatial partitioning in a widely distributed coral reef fish. Mar Biol 150:659–670

Coral Reefs (2012) 31:839–851 Langham GM (2007) Specialized avian predators repeatedly attack novel color morphs of Heliconius butterflies. Evolution 58: 2783–2787 Leray M, Beldade R, Holbrook SJ, Schmitt RJ, Planes S, Bernardi G (2010) Allopatric divergence and speciation in coral reef fish: The Three-Spot Dascyllus, Dascyllus trimaculatus, species complex. Evolution 64:1218–1230 Lessios HA (2008) The great American schism: Divergence of marine organisms after the rise of the Central American isthmus. Annual Reviews of Ecology Evolution and Systematics 39:63–91 Librado P, Rozas J (2009) DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25:1451–1452 Lin H-C, Sa´nchez-Ortiz C, Hastings PA (2009) Colour variation is incongruent with mitochondrial lineages: cryptic speciation and subsequent diversification in a Gulf of California reef fish (Teleostei: Blennioidei). Mol Ecol 18:2476–2488 McCafferty S, Bermingham E, Quenouille B, Planes S, Hoelzer G, Asoh K (2002) Historical biogeography and molecular systematics of the Indo-Pacific genus Dascyllus (Teleostei: Pomacentridae). Mol Ecol 11:1377–1392 McCartney MA, Acevedo J, Heredia C, Rico C, Quenouille B, Bermingham E, McMillan WO (2003) Genetic mosaic in a marine species flock. Mol Ecol 12:2963–2973 McMillan WO, Palumbi SR (1995) Concordant evolutionary patterns among Indo-West Pacific butterflyfishes. Proc R Soc Lond B Biol Sci 260:229–236 McMillan WO, Weigt LA, Palumbi SR (1999) Color pattern evolution, assortative mating, and genetic differentiation in brightly colored butterflyfishes (Chaetodontidae). Evolution 53:247–260 Meeker ND, Hutchinson SA, Ho L, Trede NS (2007) Method for isolation of PCR-ready genomic DNA from zebrafish tissues. Biotechniques 43:610–614 Moyer JT (1990) Social and reproductive behavior of Chaetodontoplus mesoleucus (Pomacanthidae) at Bantayan Island, Philippines, with notes on pomacanthid relationships. Japan J Ichthyol 36:459–467 Moyer JT, Nakazono A (1978) Population structure, reproductive behavior, and protogynous hermaphroditism in the angelfish Centropyge interruptus at Miyake-jima, Japan. Japan J Ichthyol 25:25–39 Narum SR (2006) Beyond Bonferroni: Less conservative analyses for conservation genetics. Conserv Genet 7:783–787 Pianka ER (1978) Evolutionary ecology. Harper and Row, New York, USA Planes S, Doherty PJ (1997a) Genetic relationships of the colour morphs of Acanthochromis polyacanthus (Pomacentridae) on the northern Great Barrier Reef. Mar Biol 130:109–117 Planes S, Doherty PJ (1997b) Genetic and color interactions at a contact zone of Acanthochromis polyacanthus: A marine fish lacking pelagic larvae. Evolution 51:1232–1243 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 Posada D (2003) Using Modeltest and PAUP* to select a model of nucleotide substitution. pp. 6.5.1–6.5.14. In: Baxevanis AD, Davison DB, Page RDM, Petsko GA, Stein LD, Stormo GD (eds) Current protocols in bioinformatics. John Wiley & Sons, Inc Posada D (2008) jModelTest: Phylogenetic model averaging. Mol Biol Evol 25:253–1256 Puebla O, Bermingham E, Guichard F, Whiteman E (2007) Colour pattern as a single trait driving speciation in Hypoplectrus coral reef fishes? Proc R Soc Lond B Biol Sci 274:1265–1271 Puebla O, Bermingham E, Guichard F (2008) Population genetic analyses of Hypoplectrus coral reef fishes provide evidence that local processes are operating during the early stages of marine adaptive radiations. Mol Ecol 17:1405–1415

851 Pyle RL (1992) Rare and unusual marines: A hybrid angelfish Centropyge flavissimus x eibli. Freshw Mar Aquar 15(98–110):212 Pyle RL (2003) A systematic treatment of the reef-fish family Pomacanthidae (Pisces: Perciformes). Ph.D. Dissertation, University of Hawai’i, Honolulu, p 422 Pyle RL, Randall JE (1994) A review of hybridization in marine angelfishes (perciformes: Pomacanthidae). Environ Biol Fish 41:127–145 Ramon ML, Lobel PS, Sorenson MD (2003) Lack of mitochondrial genetic structure in hamlets (Hypoplectrus spp.): recent speciation or ongoing hybridization? Mol Ecol 12:2975–2980 Randall JE (1998) Zoogeography of shore fishes of the Indo-Pacific region. Zool Stud 37:227–268 Randall JE, Rocha LA (2009) Chaetodontoplus poliourus, a new angelfish (Percirformes: Pomacanthidae) from the tropical western Pacific. Raffles Bull Zool 57:511–520 Reece JS, Bowen BW, Joshi K, Goz V, Larson L (2010) Phylogeography of two moray eels indicates high dispersal throughout the Indo-Pacific. J Hered 101:391–402 Rocha LA (2004) Mitochondrial DNA and color pattern variation in three western Atlantic Halichoeres (Labridae), with the revalidation of two species. Copeia 2004:770–782 Rocha LA, Bowen BW (2008) Speciation in coral reef fishes. J Fish Biol 72:1101–1121 Rocha LA, Robertson DR, Roman J, Bowen BW (2005) Ecological speciation in tropical reef fishes. Proc R Soc Lond B Biol Sci 272:573–579 Rocha LA, Craig MT, Bowen BW (2007) Phylogeography and the conservation genetics of coral reef fishes. Coral Reefs 26: 501–512 Rogers AR, Harpending H (1992) Population growth makes waves in the distribution of pairwise genetic differences. Mol Biol Evol 9:552–569 Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574 Schultz JK, Pyle RL, DeMartini E, Bowen BW (2007) Genetic connectivity among color morphs and Pacific archipelagos for the flame angelfish, Centropyge loriculus. Mar Biol 151:167–175 Seehausen O, Van Alphen JJM, Witte F (1997) Cichlid fish diversity threatened by eutrophication that curbs sexual selection. Science 277:1808–1811 Seehausen O, Mayhew PJ, Van Alphen JJM (1999) Evolution of colour patterns in East African cichlid fish. J Evol Biol 12: 514–534 Seutin G, White BN, Boag PT (1991) Preservation of avian and blood tissue samples for DNA analyses. Can J Zool 69:82–92 Spalding MD, Fox HE, Allen GR, Davidson N, Ferdan˜a ZA, Finlayson M, Halpern BS, Jorge MA, Lombana A, Lourie SA, Martin KD, McManus E, Molnar J, Recchia CA, Robertson J (2007) Marine ecoregions of the world: A bioregionalization of coastal and shelf areas. Bioscience 57:573–583 Stephens M, Donnelly P (2003) A comparison of Bayesian methods for haplotype reconstruction from population genotype data. Am J Hum Genet 73:1162–1169 Swofford DL (2000) PAUP*: Phylogenetic analysis by parsimony, version 4.0. Sinauer Associates. Sunderland, MA Takeshita GY (1976) An angel hybrid. Mar Aquarist 7:27–35 Tavare´ S (1984) Line-of-descent and genealogical processes, and their applications in population genetics models. Theor Popul Biol 26:119–164 Thresher RE, Brothers EB (1985) Reproductive ecology and biogeography of Indo-West Pacific angelfishes (Pisces: Pomacanthidae). Evolution 39:878–887 Yaakub SM, Bellwood DR, Van Herwerden L (2006) A rare hybridization event in two common Caribbean wrasses (genus Halichoeres; family Labridae). Coral Reefs 26:597–602

123