Resolving cryptic diversity in the Chromis viridis species complex (Pomacentridae): a multilocus phylogenetic and morphometric reassessment of species boundaries

Abstract

The blue-green damselfish, Chromis viridis (Cuvier, 1830), is widely distributed across Indo-Pacific coral reefs and is among the most heavily traded marine ornamental fishes. Recent molecular studies have suggested the presence of cryptic diversity within the C. viridis complex, but species boundaries remain unresolved. Multilocus phylogenetic analyses incorporating three mitochondrial markers—cytochrome c oxidase subunit I (COI), cytochrome b (cytb), and the mitochondrial control region (D-loop)—and two nuclear markers, recombination-activating gene 1 (RAG1) and recombination-activating gene 2 (RAG2), were performed to assess lineage diversification within the Chromis viridis species complex. Specimens collected from the Lakshadweep Archipelago and the Gulf of Mannar, India, formed a distinct and well-supported lineage that was clearly separated from C. viridis sensu stricto and other congeners in phylogenetic reconstructions and genetic distance analyses. Multiple species-delimitation approaches consistently supported the recognition of this lineage as an independently evolving taxon, and morphometric analyses revealed corresponding differences in body proportions and meristic characters. Based on the concordance of molecular and morphological evidence, a new species, Chromis veneta sp. nov., is herein described from the Indian Ocean. The new species differs from C. viridis in having a relatively elongate body, narrower caudal peduncle, distinctive body pigmentation, yellowish to orangish distal fin margins, and a deeply forked caudal fin with elongate lobe tips. Additional diagnostic features include differences in dentition, gill-raker counts and sagittal otolith morphology. The discovery of C. veneta sp. nov. highlights underestimated diversity within the C. viridis complex and underscores the importance of integrative taxonomy in revealing cryptic reef-fish diversity.

1 Introduction

Damselfishes belong to the family Pomacentridae, which represents one of the most diverse groups of reef-associated fishes, currently comprising 438 valid species in 29 genera, with new species continuing to be described at a steady pace (Fricke et al., 2026). Within this diversity, the blue-green damselfish Chromis viridis (Cuvier, 1830) stands out as a ubiquitous and ecologically prominent Indo-Pacific reef species belonging to the species-rich subfamily Chrominae with 128 valid species (Fricke et al., 2026). Moreover, blue green damsel fishes are among the most heavily traded coral reef fishes globally and are routinely harvested in large numbers for aquarium export (Wabnitz et al., 2003; Rhyne et al., 2012). Despite its ecological prominence and economic importance, the taxonomy of Chromis viridis has historically been obscured by nomenclatural confusion and morphological similarity with congeners.

Chromis viridis was originally described as Pomacentrus viridis Cuvier, 1830, based on an unpublished illustration reproduced in Histoire naturelle des poisons (Randall et al., 1985). In the same work, Cuvier and Valenciennes (1830) also described Heliases caeruleus from a preserved specimen. Because Pomacentrus viridis lacked a preserved name-bearing type and the original description of Heliases caeruleus was brief, the identities and correct application of the names Chromis viridis and Chromis caerulea remained uncertain for more than a century. As a result, these names were used inconsistently for blue-green damselfishes throughout the Indo-Pacific, leading to considerable nomenclatural instability and confusion over species boundaries within the genus.

The long-standing nomenclatural confusion was resolved by Randall et al. (1985), who critically re-examined the historical descriptions and available type material. They designated the original illustration of Pomacentrus viridis as the lectotype of Chromis viridis, thereby stabilizing the application of the name in accordance with the International Code of Zoological Nomenclature (ICZN, 1989). Their examination of the type specimen of Heliases caeruleus further demonstrated that it is conspecific with Chromis ternatensis (Bleeker, 1856). Consequently, the name Chromis caerulea, which had been widely applied to the latter species in earlier literature, was shown to have resulted from the misidentification of H. caeruleus. This taxonomic revision firmly established Chromis viridis (Cuvier, 1830) as the valid name for the widespread Indo-Pacific species. This interpretation was subsequently followed by Randall et al. (1987) and formally recognized by the International Commission on Zoological Nomenclature (ICZN, 1989), thereby stabilizing the usage of Chromis viridis. Since then, C. viridis has been recognized as a species distinct from C. ternatensis, while its distinction from the closely related C. atripectoralis has been clarified through subsequent taxonomic and molecular studies (Welander and Schultz, 1951; Froukh and Kochzius, 2008).

Although the taxonomy of C. viridis was largely stabilized by the mid-1980s, accumulating molecular evidence from the 2000s onward challenged the view of the species as a single, pan-Indo-Pacific taxon. Early mitochondrial DNA studies revealed pronounced genetic discontinuities among populations separated by major Indo-Pacific biogeographic barriers including the Red Sea, Indian Ocean, and Coral Triangle, indicating the presence of distinct evolutionary lineages within the nominal species. Froukh and Kochzius (2008) were among the first to explicitly investigate lineage divergence in Chromis viridis, reporting reciprocally monophyletic lineages and providing evidence of deep historical separation within the species. Subsequent studies by Messmer et al. (2012) and Hubert et al. (2017) further documented the presence of distinct evolutionary lineages, reinforcing the emerging evidence for cryptic diversity in C. viridis. A major advance was provided by the integrative study of Liu et al. (2019), which combined mitochondrial cytochrome b, nuclear RAG2, and ecological niche data across the species’ geographic range. Their results confirmed two deeply divergent genetic lineages, estimated to have originated during the Pliocene approximately 3 million years ago. Despite their substantial genetic divergence, the two lineages occupied largely overlapping ecological niches, suggesting evolutionary divergence in the absence of pronounced ecological differentiation. Collectively, these studies indicate that C. viridis, long regarded as a single morphologically uniform species, comprises a complex of deeply divergent and potentially cryptic evolutionary lineages.

Chromis viridis is widely used as a model reef fish in studies of shoaling behavior, sensory ecology, larval settlement, habitat selection, population connectivity, and responses to environmental stress because of its abundance, strong association with branching corals, and extensive Indo-Pacific distribution (Smith and Warburton, 1992; Mussi et al., 2005; Lecchini et al., 2006; Ben-Tzvi et al., 2008a, Ben-Tzvi et al., 2008b; Tang et al., 2014; Ishikawa et al., 2025). Phylogeographic investigations have additionally revealed deeply divergent genetic lineages within this nominal species, suggesting the presence of cryptic diversity with important implications for ecological and evolutionary inference (Liu et al., 2019). Unrecognized cryptic lineages may obscure patterns of adaptive divergence and lead to misleading ecological or behavioural interpretations because morphologically similar but genetically distinct taxa can differ in demographic characteristics, environmental tolerances, and evolutionary histories (Bickford et al., 2007). Consequently, recognising hidden diversity within the Chromis viridis complex is essential for accurately interpreting patterns of reef fish connectivity, resilience, and responses to environmental change.

To date, comprehensive morphological investigations aimed at diagnosing cryptic lineages within the Chromis viridis complex are lacking, and it remains unclear whether these lineages constitute distinct evolutionary species or represent cryptic intraspecific variation. This uncertainty hampers accurate biodiversity assessments and constrains effective conservation planning, ecosystem-based management, and sustainable regulation of the marine aquarium trade. Integrative taxonomic approaches that combine molecular and morphological evidence are therefore essential for resolving species boundaries within this complex. In this context, the present study undertakes an integrative delimitation of cryptic lineages and a reassessment of species boundaries in the Chromis viridis complex using multilocus genetic data and morphometric analyses, resulting in the discovery of a new species.

2 Materials and methods

2.1 Specimen collection and preservation

A total of 69 blue-green damselfish specimens were collected from Agatti and Minicoy islands in the Lakshadweep Archipelago (northern Indian Ocean) and the Gulf of Mannar (eastern Indian Ocean). Sampling at Agatti Island was conducted on three occasions between October and December 2024, during which 30 specimens were collected. Specimens were collected from coral reef habitats by accompanying local fishers during routine live-bait collections, in which lift nets were employed. In the Gulf of Mannar, 25 specimens were collected with the assistance of local fishers during a single sampling effort in January 2025. Additional sampling at Minicoy Island in December 2025 yielded 14 specimens.

The collected specimens were provisionally assigned to two distinct morphotypes based on external morphology, particularly body shape and fin colouration. The first morphotype possessed a relatively deep, laterally compressed body and lacked yellow pigmentation along the distal margins of the dorsal and anal fins. The second morphotype exhibited a comparatively more elongate body and, in adults, conspicuous yellow pigmentation along the distal margins of the dorsal and anal fins. Since this pigmentation was absent in juveniles, immature specimens were provisionally assigned to the second morphotype primarily based on body shape. At Agatti Island, 13 of the 30 specimens belonged to the second morphotype, whereas 14 of the 25 specimens collected from the Gulf of Mannar were assigned to this morphotype. In contrast, all 14 specimens collected from the sampling site at Minicoy Island belonged to the first morphotype, and no individuals of the second morphotype were recorded at that locality.

Both morphotypes were collected sympatrically at Agatti Island and in the Gulf of Mannar, indicating their coexistence within the same reef habitats. Their sympatric occurrence was further documented through in situ underwater photographs from Agatti Island, where both morphotypes were observed together on the same reef. All specimens were photographed alive, either in situ or immediately after capture, to document live colouration before preservation.

Following photography, all specimens were euthanized by immersion in buffered MS-222 (tricaine methanesulfonate; 0.05% w/v), with sodium bicarbonate added to neutralize the acidity of the anaesthetic solution, following the guidelines of the American Fisheries Society and the American Veterinary Medical Association. Specimens were maintained in the solution until opercular (gill) movements had ceased and for at least an additional 10 min to ensure death. Prior to fixation, muscle and fin tissue samples were excised and preserved in 95–100% ethanol for subsequent molecular analyses. Voucher specimens were then fixed in 10% neutral-buffered formalin for 24–48 h, thoroughly rinsed with water, and transferred to 70% ethanol for long-term preservation.

Subsequent integrative analyses incorporating morphometric, meristic, and multilocus molecular data confirmed that the two morphotypes represent Chromis viridis (Cuvier, 1830) and the proposed new species, respectively.

2.2 DNA extraction, amplification, and sequencing

Genomic DNA was isolated from fin clips or muscle tissue using the standard phenol–chloroform extraction method described by Sambrook and Russell (2001). DNA concentration and purity were evaluated using a NanoDrop spectrophotometer (Thermo Fisher Scientific), and DNA integrity was confirmed by electrophoresis on a 0.8% agarose gel. Partial fragments of three mitochondrial markers—cytochrome c oxidase subunit I (COI), cytochrome b (Cytb), and the Mitochondrial control region (D-loop)—as well as two nuclear markers, recombination-activating gene 1 (RAG1) and recombination-activating gene 2 (RAG2), were amplified by polymerase chain reaction (PCR) using the primer pairs listed in Table 1.

Amplification of the COI fragment was performed with an initial denaturation at 95 °C for 2 min, followed by 35 cycles of denaturation at 94 °C for 30 s, annealing at 50 °C for 40 s, and extension at 72 °C for 40 s, with a final extension at 72 °C for 10 min. The Cytb fragment was amplified using an initial denaturation at 94 °C for 2 min, followed by 41 cycles of denaturation at 94 °C for 30 s, annealing at 57 °C for 30 s, and extension at 72 °C for 40 s, with a final extension at 72 °C for 2 min (Palumbi, 1996). Amplification of the mitochondrial control region (D-loop) followed the protocol of Messmer et al. (2012) with minor modifications, consisting of an initial denaturation at 95 °C for 5 min, followed by 35 cycles of denaturation at 94 °C for 45 s, annealing at 50 °C for 45 s, and extension at 72 °C for 45 s, with a final extension at 72 °C for 5 min. PCR amplification of the nuclear genes (RAG1 and RAG2) was performed following the protocol of Quenouille et al. (2004).

All PCR reactions were performed in a final volume of 25 μL containing 12.5 μL of 2× Takara EmeraldAmp Max PCR Master Mix, 10.5 μL of nuclease-free water, 0.5 μL each of forward and reverse primers (10 μM), and 1 μL of genomic DNA template. Amplified products were maintained at 4 °C and verified by electrophoresis on a 1% agarose gel. Successful PCR products were purified and sequenced bidirectionally on an ABI 3730XL automated DNA sequencer (Applied Biosystems, USA) at GENESPEC, Kochi, Kerala, India.

2.3 Sequence alignment and genetic dataset assembly

The quality of the initial sequences was assessed using ABI Sequence Scanner (Applied Biosystems), and low-quality reads were excluded based on chromatogram inspection. Forward and reverse sequence reads were assembled, manually checked, and edited using the software BioEdit (Hall, 1999) to generate high-quality consensus sequences.

2.4 Phylogenetic analyses

Phylogenetic relationships among lineages of the Chromis viridis species complex were inferred using partial mitochondrial (COI, Cytb and D-loop) and nuclear (RAG1 and RAG2) gene sequences under both Maximum Likelihood (ML) and Bayesian Inference (BI) frameworks. For the ML analyses, the best-fitting nucleotide substitution model for each dataset was selected using the ModelFinder Plus (MFP) algorithm implemented in IQ-TREE v2 (Minh et al., 2020) based on the Bayesian Information Criterion (BIC) (Kalyaanamoorthy et al., 2017). Maximum Likelihood phylogenetic trees were reconstructed in IQ-TREE, and nodal support was assessed using 1,000 non-parametric bootstrap replicates. Bayesian phylogenetic analyses were performed in MrBayes v3.2 (Ronquist et al., 2012). The best-fitting nucleotide substitution model for each dataset was identified using MrModeltest v2.3 (Nylander, 2004), and the selected model was implemented in MrBayes. Two independent Markov chain Monte Carlo (MCMC) analyses, each consisting of four chains (one cold and three heated), were run for 15 million generations, with trees sampled every 1,000 generations. The first 10% of sampled trees were discarded as burn-in, and the remaining trees were used to construct a majority-rule consensus tree and estimate Bayesian posterior probabilities. Convergence and mixing of the MCMC chains were assessed using Tracer v1.7 (Rambaut et al., 2018). Analyses were considered to have converged when all estimated parameters exhibited effective sample size (ESS) values greater than 200, indicating adequate sampling of the posterior distribution. For the concatenated dataset, Maximum Likelihood (ML) analysis was performed in IQ-TREE using an initial partitioning scheme of 13 partitions, corresponding to the three codon positions of each protein-coding gene and one partition for the non-coding D-loop. The optimal partitioning scheme and substitution models were selected using ModelFinder with the MFP+MERGE option. Branch support was assessed using 1,000 SH-aLRT and 1,000 ultrafast bootstrap (UFBoot) replicates with the BNNI correction. For Bayesian analysis, the concatenated alignment was partitioned by gene, and an independent nucleotide-substitution model was assigned to each partition according to the selected best-fit models. Model parameters were unlinked among partitions, allowing each gene to have independent nucleotide frequencies, substitution rates, and among-site rate variation. To provide phylogenetic context, representatives of the Hoplochromis species group (Fowler and Bean, 1928), including Chromis viridis, C. atripectoralisWelander and Schultz, 1951, C. flavaxillaRandall, 1994, and C. ternatensis (Bleeker, 1856), were included in the analyses according to sequence availability in the NCBI GenBank database. The trees were rooted with appropriate out group taxa and visualized in iTOL v7 (Letunic and Bork, 2024).

2.5 Genetic distance

Genetic distance analyses were performed to quantify levels of sequence variation within and between the morphologically recognized species and to evaluate the genetic distinctiveness of the proposed new species relative to its closest congeners. Pairwise sequence divergences were estimated independently for each molecular marker (COI, Cytb, D-loop, RAG1, and RAG2). Uncorrected pairwise genetic distances (p-distances) were calculated in MEGA 12 (Kumar et al., 2024) using the Uniform Rates substitution model with pairwise deletion of missing data. Mean intra- and interspecific genetic distances were computed for all specimens included in the analyses. Additionally, pairwise genetic distances were estimated between the proposed new species and representative members of the Chromis (Hoplochromis) species group available in GenBank, including C. viridis, C. atripectoralis, C. ternatensis, and C. flavaxilla, to assess the extent of genetic divergence among closely related taxa.

2.6 Species delimitation analyses

Species boundaries within the Chromis viridis complex were evaluated using complementary distance-based, tree-based, and multispecies coalescent approaches. Congruence among the different delimitation methods and phylogenetic analyses was considered evidence supporting independently evolving evolutionary lineages.

2.6.1 Automatic barcode gap discovery

Species delimitation using the Automatic Barcode Gap Discovery (ABGD) method was performed separately for each of the four loci (COI, Cytb, D-loop, RAG1 and RAG2) using the Windows executable version of ABGD downloaded from the iTaxoTools website (https://itaxotools.org/download.html). Pairwise genetic distances were calculated under the Kimura two-parameter (K2P) substitution model. Both the initial and recursive partitions were examined, and the partition showing the greatest congruence with the phylogenetic analyses was considered the primary species hypothesis (Puillandre et al., 2012).

2.6.2 Assemble species by automatic partitioning

Species delimitation was further assessed using Assemble Species by Automatic Partitioning (ASAP) through Windows executable version of ASAP downloaded from the iTaxoTools website (https://itaxotools.org/download.html). Each of the four loci (COI, Cytb, D-loop, RAG1 and RAG2) was analysed independently using Kimura two-parameter (K2P) genetic distances with the default parameters. The partition with the lowest ASAP score was considered the primary species hypothesis, while other top-ranked partitions were also examined to evaluate the consistency of species boundaries among loci (Puillandre et al., 2021).

2.6.3 Multi-rate poisson tree processes

Tree-based species delimitation was conducted using the multi-rate Poisson Tree Processes (mPTP) method (Kapli et al., 2017). Separate analyses were performed for each of the four loci (COI, Cytb, D-loop, RAG1 and RAG2). Maximum-likelihood (ML) phylogenetic trees were reconstructed independently for each gene using IQ-TREE and used as input trees for mPTP species delimitation analyses. The analyses were performed using the mPTP web server (https://mptp.h-its.org/) under the maximum-likelihood framework with default parameters. Species hypotheses recovered consistently across multiple loci were considered more robust than those supported by a single marker.

2.6.4 Bayesian phylogenetics and phylogeography

Multilocus species delimitation was performed in BPP v4.4.0 under the multispecies coalescent (MSC) model (Yang and Rannala, 2010) using three mitochondrial loci (COI, cytb, and D-loop) and two nuclear loci (RAG1 and RAG2). Species delimitation was evaluated using both analysis A10, which estimates species boundaries on a fixed guide tree (Yang and Rannala, 2010; Rannala and Yang, 2013), and analysis A11, which jointly infers species delimitation and species-tree topology (Yang and Rannala, 2014). The guide tree for A10 was inferred from the concatenated multilocus maximum-likelihood and Bayesian phylogenetic analyses. Two independent Markov chain Monte Carlo (MCMC) runs with different random seeds were performed for each analysis to assess convergence. Population size (θ) and root divergence time (τ₀) priors were assigned inverse-gamma distributions following the recommendations of Yang and Rannala for shallow species-level divergences, whereas internal divergence times were estimated using the Dirichlet prior implemented in BPP (Yang and Rannala, 2010; Yang, 2015). Species delimitations with posterior probabilities (PP) ≥ 0.95 were considered strongly supported.

2.7 Morphometric analysis

A total of 10 specimens of the proposed new species, comprising one holotype (NBFGR/POMCVEN) and nine paratypes (NBFGR/POMCVEN.1–POMCVEN.9), were examined for the formal species description and deposited in the National Fish Germplasm Repository, recognized repository for fish germplasm, http://nbaindia.org/ uploaded/pdf/Repositorites_of_BDAct.pdf. Ten specimens of Chromis viridis representing the comparative material were also deposited in the repository, including five specimens from Agatti Island (NBFGR/POMCVIR.1–POMCVIR.5) and five from Minicoy Island (NBFGR/POMCVIR.6–POMCVIR.10). The identity of the C. viridis specimens was confirmed by their agreement with published morphological descriptions (Randall et al., 1981; Welander and Schulz, 1951) and by molecular analyses, which assigned all specimens to Clade A of the Chromis viridis species complex (Liu et al., 2019).

Morphometric measurements for the formal species description were taken following Allen and Allen (2021) and Pyle et al. (2008). Measurements were recorded to the nearest 0.01 mm using digital calipers (Mitutoyo, Japan) and are presented to one decimal place. Meristic counts were made under a stereomicroscope (Ivesta 3, Leica Microsystems, Germany). In the species description, values separated by a vertical bar (|) represent measurements taken from the left and right sides of the specimen, respectively. Vertebral counts were determined using high-resolution micro-computed tomography (Bruker SkyScan 1276, Belgium). Counts included all vertebrae from the first vertebra articulating with the cranium to the terminal vertebra immediately anterior to the hypural plate. Fresh colouration was documented by standardized digital photography immediately after collection, and additional photographs of live individuals were obtained in situ to record natural colour patterns. Colouration in preservative was recorded after fixation in 10% buffered formalin followed by transfer to 70% ethanol.

Comparative morphometric analyses were conducted on 40 specimens, comprising 20 specimens of Chromis viridis (10 from Agatti Island and 10 from Minicoy Island) and 20 specimens of the proposed new species (11 from Agatti Island and 9 from the Gulf of Mannar). The dataset comprised the 10 Chromis viridis comparative specimens (CV1 – CV10), the 10 type specimens (CVEN1 – CVEN10; holotype and paratypes) of the proposed new species used for its formal description, and an additional 10 non-type specimens of each species for morphometric analyses. Thirty-three morphometric characters were recorded for each specimen. To minimize the influence of ontogenetic variation, only individuals larger than 40 mm standard length (SL) were included in the multivariate analyses. To remove the effect of body size and facilitate shape-based comparisons among specimens, morphometric measurements were size-adjusted using the allometric transformation of Elliott et al. (1995):

where M is the original measurement, M_adj is the size-adjusted measurement, L_o is the standard length of an individual specimen, L_s is the overall mean standard length of all specimens, and b is the allometric coefficient estimated from the slope of the regression of log-transformed measurements against log-transformed standard length. The standardized, size-adjusted dataset was then used for all subsequent statistical analyses.

Differences in individual morphometric characters between the two morphotypes were assessed using one-way analysis of variance (ANOVA). Overall differences in morphometric profiles were evaluated using multivariate analysis of variance (MANOVA), with species as the grouping factor and all morphometric variables analysed simultaneously. Statistical significance of the multivariate model was assessed using four standard multivariate test statistics: Wilks’ lambda, Pillai’s Trace, Hotelling–Lawley Trace, and Roy’s Largest Root (Johnson and Wichern, 2007). Principal component analysis (PCA) was performed on the standardized, size-adjusted morphometric dataset to summarize patterns of morphological variation and visualize phenotypic differentiation between species. The contribution of each morphometric variable to the principal components was assessed from the corresponding loading vectors (Jolliffe, 2002).

Linear discriminant analysis (LDA) was performed to identify the combination of morphometric variables that best discriminated between species. Model performance was evaluated using leave-one-out cross-validation (LOOCV), and standardized discriminant coefficients were examined to identify the variables contributing most strongly to species separation (Fisher, 1936). Overall morphological differentiation was further assessed using permutational multivariate analysis of variance (PERMANOVA) based on Euclidean distance matrices with 9,999 permutations, and homogeneity of multivariate dispersions was evaluated using PERMDISP to distinguish differences in group centroids from differences in within-group dispersion (Anderson, 2001, 2006). Statistical significance was assessed at α = 0.05. All statistical analyses were performed in R version 4.5.0 (R Core Team, 2025).

3 Results

3.1 Sequence similarity searches (nBLAST)

BLASTn searches of the mitochondrial (COI, Cytb, and D-loop) and nuclear (RAG1 and RAG2) sequences against the NCBI GenBank database consistently identified specimens of Chromis viridis as C. viridis across all loci. In contrast, sequences of the proposed new species exhibited locus-dependent matches. For the Cytb and RAG2 datasets, the highest sequence similarities corresponded to Clade B of the Chromis viridis species complex reported by Liu et al. (2019). Likewise, D-loop sequences showed the highest similarity to the cryptic mitochondrial lineage within the C. viridis species complex reported by Messmer et al. (2012).

For the COI dataset, sequences of the proposed new species showed the highest similarity to GenBank accessions identified as Chromis caerulea (FJ583146–FJ583148) and Chromis sp. (OL410087). The C. caerulea accessions represent specimens identified under a name that was historically misapplied to the Indo-Pacific blue-green damselfish complex (Randall et al., 1985).

3.2 Phylogenetic analysis

Phylogenetic analyses were conducted using mitochondrial (COI, Cytb, and D-loop) and nuclear (RAG1 and RAG2) gene datasets. Details of the newly generated sequences and GenBank reference sequences included in the analyses are provided in Tables 2 and 3, respectively. Phylogenetic trees were rooted using appropriate outgroup taxa. The COI and cytb phylogenies were rooted using Amphiprion percula (OP918111 and KF264297, respectively) and A. nigripes (OQ692056 and KF264292, respectively) as outgroup taxa. The D-loop phylogeny was rooted using Chromis chromis (EF489720) and C. limbata (KT844444) as outgroup taxa, whereas the RAG1 and RAG2 phylogenies were rooted using Dascyllus trimaculatus (AY208655 and AY279878 respectively) and D. aruanus (AY208649 and FJ616778 respectively).

3.2.1 COI dataset

The COI dataset comprised 46 sequences with an aligned length of 539 bp, of which 383 sites (71.1%) were conserved, 156 (28.9%) were variable, and 134 (24.9%) were parsimony informative. ModelFinder implemented in IQ-TREE selected TPM2u+F+G4 as the best-fitting substitution model under the Bayesian Information Criterion (BIC), whereas MrModeltest identified GTR+I as the optimal model under the Akaike Information Criterion (AIC) for Bayesian inference. Bayesian analyses of the COI dataset showed satisfactory convergence, with all estimated parameters exhibiting effective sample size (ESS) values greater than 200. The log-likelihood (ESS = 2031), posterior probability (ESS = 2702), tree length (ESS = 2702), substitution rate parameters (ESS = 1408–2234), nucleotide base frequencies (ESS = 1469–2434), gamma shape parameter (ESS = 788), and proportion of invariant sites (ESS = 507) all indicated adequate mixing and sampling of the posterior distribution. Maximum Likelihood (ML) and Bayesian Inference (BI) analyses recovered identical topologies. Four reciprocally monophyletic species-level clades were resolved within the Hoplochromis species group, corresponding to Chromis flavaxilla, C. ternatensis, C. atripectoralis, C. veneta sp. nov., and C. viridis (Figure 1). All newly generated specimens of C. veneta sp. nov. formed a strongly supported monophyletic lineage (BS = 100; BPP = 1.00) together with GenBank sequences FJ583146–FJ583148 and OL410087. The C. viridis clade comprised all newly generated sequences together with published reference sequences. Chromis veneta sp. nov. and C. viridis were recovered as sister taxa with strong support (BS = 77; BPP = 0.95).

Figure 1

3.2.2 Cytb dataset

The cytochrome b (cytb) dataset comprised 38 sequences with a final alignment length of 911 bp, of which 670 sites (73.5%) were conserved, 241 (26.5%) were variable, and 172 (18.9%) were parsimony informative. ModelFinder implemented in IQ-TREE selected TN+F+I as the best-fitting nucleotide substitution model under the Bayesian Information Criterion (BIC), whereas MrModeltest identified GTR+I as the optimal model under the Akaike Information Criterion (AIC) for Bayesian inference. Bayesian analyses of the Cytb dataset showed satisfactory convergence, with all estimated parameters exhibiting ESS values greater than 200. The log-likelihood (ESS = 2359), posterior probability (ESS = 2088), tree length (ESS = 2098), substitution rate parameters (ESS = 1710–2042), nucleotide base frequencies (ESS = 2057–2661), and proportion of invariant sites (ESS = 2598) indicated adequate mixing and sampling of the posterior distribution. Maximum Likelihood (ML) and Bayesian Inference (BI) analyses recovered identical topologies, resolving three well-supported species-level clades corresponding to Chromis veneta sp. nov., C. atripectoralis, and C. viridis (Figure 2). All major nodes received maximal support (BS = 100; BPP = 1.00). The focal species, Chromis veneta sp. nov., formed a strongly supported monophyletic lineage (BS = 100; BPP = 1.00) comprising all newly generated sequences together with previously published reference sequences. This clade was recovered as the sister lineage to the clade containing C. atripectoralis and C. viridis. Chromis atripectoralis was recovered as a distinct monophyletic clade (BS = 100; BPP = 1.00) and was resolved as the sister species of C. viridis. The C. viridis clade included all newly generated specimens together with published reference sequences and was also recovered with maximal nodal support (BS = 100; BPP = 1.00).

Figure 2

3.2.3 Mitochondrial control region (D-loop) dataset

The D-loop dataset comprised 37 sequences with a final alignment length of 389 bp, of which 204 sites (52.44%) were conserved, 182 (46.8%) were variable, and 131 (33.7%) were parsimony-informative. ModelFinder implemented in IQ-TREE selected K3Pu+F+G4 as the best-fitting nucleotide substitution model under the Bayesian Information Criterion (BIC), whereas MrModeltest selected GTR+G under the Akaike Information Criterion (AIC) for Bayesian inference. Bayesian MCMC analyses converged satisfactorily, with ESS values exceeding 1,600 for all estimated parameters, including the log-likelihood (ESS = 2342), posterior probability (ESS = 2565), tree length (ESS = 2567), substitution-rate parameters (ESS = 1694–2166), nucleotide base frequencies (ESS = 1723–2106), and the gamma shape parameter (ESS = 2702). Maximum Likelihood (ML) and Bayesian Inference (BI) analyses recovered identical topologies, resolving three reciprocally monophyletic species-level clades corresponding to Chromis veneta sp. nov., C. atripectoralis, and C. viridis (Figure 2). The focal species, C. veneta sp. nov., formed a strongly supported monophyletic lineage (BS = 97; BPP = 0.99) comprising all newly generated specimens together with previously published reference sequences (JX231615–JX231620). Chromis atripectoralis and C. viridis each formed distinct monophyletic clades with maximal support (BS = 100; BPP = 1.00). The sister relationship between these two species was also strongly supported (BS = 100; BPP = 1.00).

3.2.4 RAG1 dataset

The partial RAG1 dataset comprised 26 sequences with an aligned length of 973 bp, of which 902 (92.7%) sites were conserved, 71 (7.3%) were variable, and 52 (5.3%) were parsimony informative. The best-fit substitution models were TPM2+R3 for ML (IQ-TREE) and K80+I+G for BI (MrBayes). Bayesian analyses converged satisfactorily, with all estimated sample size (ESS) values exceeding 200, indicating adequate mixing and convergence of the Markov chain Monte Carlo (MCMC) runs (LnL = 2702, LnPr = 2647, Tree Length = 2665, κ = 2596, α = 1978, proportion of invariant sites = 1516).

The ML and BI analyses recovered identical topologies, resolving three monophyletic species-level clades corresponding to Chromis atripectoralis, C. veneta, and C. viridis (Figure 2). Chromis atripectoralis formed a strongly supported clade (100/1.00). Chromis veneta was recovered as sister to C. atripectoralis with moderate support (62/0.65), whereas C. viridis formed a distinct, well-supported clade (90/0.96) containing all newly generated sequences together with the GenBank C. viridis reference sequences (AY208635.1 and MW631409.1). The separation between the C. veneta and C. viridis lineages received strong support (100/1.00).

3.2.5 RAG2 dataset

The RAG2 dataset comprised 33 sequences with a final alignment length of 705 bp, of which 627 sites (88.9%) were conserved, 78 (11.1%) were variable, and 53 (7.5%) were parsimony informative. ModelFinder implemented in IQ-TREE selected K2P+I as the best-fitting nucleotide substitution model under the Bayesian Information Criterion (BIC), whereas MrModeltest selected SYM+I+G under the Akaike Information Criterion (AIC) for Bayesian inference. Bayesian MCMC analyses converged satisfactorily, with all estimated parameters exhibiting effective sample size (ESS) values well above the recommended threshold of 200. The log-likelihood (ESS = 2461), posterior probability (ESS = 2702), and tree length (ESS = 2702) showed adequate mixing. ESS values for the substitution-rate parameters ranged from 1756 to 2220, while the gamma shape parameter and proportion of invariant sites had ESS values of 2263 and 1625, respectively. Maximum Likelihood (ML) and Bayesian Inference (BI) analyses recovered largely congruent species-level topologies, resolving three reciprocally monophyletic lineages corresponding to Chromis veneta sp. nov., C. viridis, and C. atripectoralis (Figure 2). All newly generated C. veneta sp. nov. sequences, together with published reference sequences (MH743493, MH743577, MH743570, MH743596 and MH743547), formed a distinct monophyletic lineage. All newly generated C. viridis sequences clustered with published reference sequences (MH743616, MH743668, MH743677, MH743437 and MH743601) in a separate monophyletic clade.

The RAG2 topology differed from those inferred from the mitochondrial markers (COI, cytochrome b, and D-loop), in which C. veneta sp. nov. was consistently recovered as sister to C. viridis. In contrast, the nuclear RAG2 marker recovered C. veneta sp. nov. as sister to C. atripectoralis, while C. viridis occupied the basal position within the ingroup. Despite this topological incongruence, all analyses consistently supported the reciprocal monophyly of the three species.

3.2.6 Concatenated dataset

The mitochondrial COI (539 bp), Cyt b (911 bp), and D-loop (389 bp), along with the nuclear RAG1 (973 bp) and RAG2 (705 bp) loci, were concatenated to generate a multilocus dataset. Although the combined raw sequence length was 3,517 bp, the final concatenated alignment comprised 23 taxa and 3,677 aligned characters, including gap-containing positions retained during multiple sequence alignment. ModelFinder optimized the initial 13 partitions into eight partitions, each assigned an independent best-fitting substitution model. The optimized analysis merged partitions with similar evolutionary patterns, resulting in eight partitions, each assigned an independent best-fitting substitution model. The selected models were HKY+F+I, HKY+F, TN+F, TIM+F, HKY+F+I+R2, TPM3u+F+R2, K2P+I, and TPM2u+F+R2. The D-loop partition exhibited the highest level of phylogenetically informative variation, whereas the nuclear loci were comparatively conserved. The maximum-likelihood analysis recovered a strongly supported ingroup clade relative to the outgroups Dascyllus trimaculatus and D. aruanus (SH-aLRT/UFBoot = 100/100), with a final log-likelihood value of −9750.0303. Chromis viridis formed a highly supported monophyletic lineage (99.6/100), while all specimens of Chromis veneta sp. nov. clustered into a distinct clade with maximum support (100/100). For the Bayesian inference, the alignment was partitioned by gene, with COI, Cytb, D-loop, RAG1, and RAG2 comprising 539, 911, 549, 973, and 705 bp, respectively. Partition-specific substitution models were implemented as follows: GTR + Γ for COI and D-loop, GTR + I for Cytb, K80 + I for RAG1, and GTR + I + Γ for RAG2. Model parameters, including nucleotide frequencies, relative substitution rates, gamma shape parameters, and the proportion of invariant sites, were unlinked across partitions. Relative evolutionary rates among partitions were allowed to vary. Bayesian inference based on the concatenated multilocus dataset recovered three distinct species-level lineages. Chromis viridis specimens formed a monophyletic lineage with maximum posterior probability support (PP = 1.00). Chromis atripectoralis was recovered as a distinct lineage, sister to the C. veneta sp. nov. clade, which also formed a well-supported monophyletic lineage (PP = 1.00). The analysis therefore strongly supported the genetic distinctiveness and reciprocal monophyly of C. viridis and C. veneta sp. nov. (Figure 3).

Figure 3

3.3 Genetic distance

The average uncorrected p-distance analysis based on mitochondrial COI sequences revealed pronounced genetic divergence between Chromis veneta sp. nov. and its congeners (Table 4). The new species exhibited a mean genetic divergence of 7.1% from Chromis viridis. In contrast, negligible genetic divergence (0.0–0.4%) was observed between C. veneta sp. nov., sequences identified as C. caerulea (GenBank accessions FJ583148.1, FJ583146.1, and FJ583147.1), and Chromis sp. (OL410087.1), indicating that these sequences likely represent the same evolutionary lineage. Mean genetic divergence between C. veneta sp. nov. and C. atripectoralis was also 7.1%, whereas substantially higher divergences were observed relative to C. ternatensis (15.4%) and C. flavaxilla (15.8%).

The mean uncorrected p-distance analysis based on the mitochondrial Cytb dataset (Table 4) revealed substantial genetic divergence among lineages within the Chromis viridis species complex. Chromis veneta sp. nov. exhibited a mean genetic divergence of 9.3% from C. viridis and 9.6% from C. atripectoralis. Sequences previously assigned to Chromis viridis Clade A by Liu et al. (2019) showed negligible divergence from C. viridis (0.4%) but were highly divergent from C. veneta sp. nov. (9.1%). In contrast, Clade B sequences exhibited substantial divergence from C. viridis (9.7%) but were nearly identical to C. veneta sp. nov., differing by only 0.3%. These patterns strongly support the correspondence of the previously recognised Clade B lineage with C. veneta sp. nov., while confirming the genetic distinctiveness of Clade A as C. viridis.

The mitochondrial control region (D-loop) exhibited the highest levels of genetic divergence among the markers examined (Table 4). Mean uncorrected p-distance between Chromis veneta sp. nov. and C. viridis was 19.0%, increasing to 20.4% between C. veneta sp. nov. and C. atripectoralis. Sequences JX231629.1, JX231634.1, and JX231644.1 showed relatively low genetic divergence from C. viridis (1.9–2.4%) but substantial divergence from C. veneta sp. nov. (22.3–22.7%), indicating their affinity with C. viridis. Conversely, sequences JX231616.1, JX231615.1, and JX231620.1 exhibited high divergence from C. viridis (21.6–22.4%) but low divergence from C. veneta sp. nov. (1.1–1.6%), demonstrating that these sequences are conspecific with C. veneta sp. nov. rather than C. viridis.

Mean intraspecific RAG1 genetic distances were low, ranging from 0.0000 in Chromis atripectoralis to 0.0053 ± 0.003 (0.53%) in C. viridis and 0.0077 ± 0.008 (0.77%) in C. veneta (Table 4). Mean interspecific distances exceeded intraspecific values, with C. veneta differing from C. viridis by 0.0117 ± 0.001 (1.17%), from C. atripectoralis by 0.0154 ± 0.002 (1.54%), and C. viridis differing from C. atripectoralis by 0.0164 ± 0.002 (1.64%). Although sequence divergence was low, as expected for the slowly evolving nuclear RAG1 marker, the consistently higher interspecific than intraspecific distances support the genetic distinctiveness of C. veneta, C. viridis, and C. atripectoralis.

The nuclear RAG2 dataset exhibited lower levels of sequence divergence than the mitochondrial markers but consistently resolved Chromis veneta sp. nov. as a distinct genetic lineage (Table 4). The mean uncorrected p-distance between C. veneta sp. nov. and C. viridis was 1.7%, increasing to 2.1% relative to C. atripectoralis and 5.7% relative to C. ternatensis. Intraspecific divergence was low in both C. veneta sp. nov. (0.5 ± 0.3%) and C. viridis (0.6 ± 0.3%). Despite the comparatively conserved nature of the RAG2 marker, the observed interspecific divergence exceeded intraspecific variation, providing independent nuclear evidence supporting the genetic distinctiveness of C. veneta sp. nov. from its congeners.

The complete set of uncorrected pairwise sequence divergences (p-distances) for all analysed sequences across the four molecular markers is provided in the Supplementary Material (Supplementary Table 1).

3.4 Species delimitation

Distance-based species delimitation was performed independently for each locus using Automatic Barcode Gap Discovery (ABGD) and Assemble Species by Automatic Partitioning (ASAP), whereas tree-based species delimitation was conducted using the multi-rate Poisson Tree Processes (mPTP) method (Figure 1). The COI dataset comprised 23 sequences, including 20 newly generated sequences (10 Chromis veneta sp. nov. and 10 C. viridis) and three GenBank reference sequences representing C. atripectoralis (n = 1), C. ternatensis (n = 1), and C. flavaxilla (n = 1). The Cytb, D-loop and RAG1 datasets each consisted of 21 sequences, including the same 20 newly generated sequences and one C. atripectoralis reference sequence, whereas the RAG2 dataset comprised 22 sequences with additional reference sequences of C. atripectoralis and C. ternatensis. Bayesian species delimitation was subsequently performed in BPP under the multispecies coalescent (MSC) model using a standardized four-locus dataset (COI, Cytb, D-loop, RAG1 and RAG2) comprising 22 individuals (10 Chromis veneta sp. nov., 10 C. viridis, one C. atripectoralis, and one C. ternatensis) to minimize missing loci across markers and enable joint estimation of species boundaries and species relationships.

3.4.1 ABGD

Automatic Barcode Gap Discovery (ABGD) analysis of the COI dataset recovered a stable partition of five Molecular Operational Taxonomic Units (MOTUs) across a broad range of prior intraspecific divergence values (P = 0.004642–0.059948) (Figure 1), corresponding to Chromis veneta sp. nov., C. viridis, C. atripectoralis, C. ternatensis, and C. flavaxilla. The recovery of C. veneta sp. nov. as a distinct MOTU provides strong support for its recognition as an independent evolutionary lineage. Similarly, ABGD analyses of the Cytb and D-loop datasets recovered stable partitions of three MOTUs (P = 0.002783–0.059948 and P = 0.021544–0.100000, respectively), consistently resolving C. veneta sp. nov., C. viridis, and C. atripectoralis as distinct genetic lineages. Although the D-loop dataset transiently split C. veneta sp. nov. at lower prior values, recursive partitioning recovered a single, well-defined MOTU. In contrast, the conserved nuclear RAG2 marker showed lower resolution, recovering C. veneta sp. nov., C. viridis + C. atripectoralis, and C. ternatensis as three MOTUs at lower prior values (P = 0.001–0.002783), while merging C. veneta sp. nov., C. viridis, and C. atripectoralis into a single MOTU at P ≥ 0.004642, reflecting the limited species-level resolving power of this locus. This reduced discriminatory power was even more pronounced in the RAG1 dataset, where ABGD consistently recovered a single MOTU comprising all Chromis sequences across all five recursive partitions, with no subdivision corresponding to C. veneta sp. nov., C. viridis, or C. atripectoralis.

3.4.2 ASAP

ASAP analysis using the Kimura two-parameter (K2P) distance model recovered a five-species partition for the COI dataset (ASAP score = 4.5), corresponding to Chromis veneta sp. nov., C. viridis, C. atripectoralis, C. ternatensis, and C. flavaxilla. The Cytb and D-loop datasets each recovered three species-level clusters (ASAP scores = 1.0 and 1.5, respectively), consistently delimiting C. veneta sp. nov., C. viridis, and C. atripectoralis (Figure 1). The ASAP analysis of the RAG1 dataset identified a three-MOTU partition (ASAP score = 8.5) that was consistent with the existing taxonomy. However, all Chromis sequences were assigned to a single MOTU (ASAP score = 1), with no subdivision corresponding to C. veneta, C. viridis, or C. atripectoralis, indicating that the RAG1 marker lacks sufficient variation for distance-based species delimitation within the genus. For the RAG2 dataset, a four-species partition (ASAP score = 4.0) was selected, recognizing C. veneta sp. nov., C. viridis, C. atripectoralis, and C. ternatensis as distinct lineages. Although the lowest ASAP score (1.0) supported a two-species partition, it merged the well-established species C. viridis and C. atripectoralis into a single cluster and was therefore rejected in favour of the partition congruent with phylogenetic, multilocus species delimitation and morphological evidence.

3.4.3 mPTP

The mPTP analysis of the COI dataset recovered five species-level lineages corresponding to Chromis veneta sp. nov., C. viridis, C. atripectoralis, C. ternatensis, and C. flavaxilla, fully concordant with the current taxonomy. Similarly, the cytb dataset recovered three Molecular Operational Taxonomic Units (MOTUs) corresponding to C. veneta sp. nov., C. viridis, and C. atripectoralis, with all species receiving maximal support (PP = 1.000) and no evidence of intraspecific lineage splitting (Figure 1). The D-loop dataset likewise recovered three species-level lineages corresponding to C. veneta sp. nov., C. viridis, and C. atripectoralis, whereas the RAG2 dataset resolved four lineages corresponding to C. veneta sp. nov., C. viridis, C. atripectoralis, and C. ternatensis. The RAG1 dataset recovered three species-level lineages corresponding to C. veneta, C. viridis, and C. atripectoralis. All specimens of each species were assigned to a single, well-supported MOTU, and species delimitation received maximal support (PP = 1.000), with no evidence of intraspecific lineage splitting.

3.4.4 BPP

Bayesian species delimitation under the multispecies coalescent strongly supported the recognition of Chromis veneta sp. nov. as a distinct evolutionary lineage. In the A10 analysis using a fixed guide tree, the fully resolved four-species delimitation model received higher posterior support (PP = 0.9969), recognizing Chromis veneta sp. nov. as a distinct species separate from C. viridis, C. atripectoralis, and C. ternatensis. The posterior probabilities for the internal nodes of the guide tree were high (PP = 0.9969–1.0000), supporting the independent evolutionary status of all four species. In the joint species delimitation and species-tree analysis (A11), the four-species model was also strongly supported (PP = 0.99898), whereas the alternative three-species model, which merged C. atripectoralis and C. ternatensis, received only PP = 0.00102. At the species level, C. veneta sp. nov. and C. viridis were delimited with maximal posterior support (PP = 1.0000), while C. atripectoralis and C. ternatensis were each supported as distinct species (PP = 0.99898) (Figure 1).

3.4.5 Consensus species delimitation

The integrative species delimitation analyses consistently supported Chromis veneta sp. nov. as a distinct evolutionary lineage within the C. viridis species complex. The mitochondrial markers (COI, cytb, and D-loop) marker showed concordant species boundaries across ABGD, ASAP, mPTP, and BPP analyses, whereas the more conserved nuclear RAG1 and RAG2 markers resolved C. veneta sp. nov. in phylogenetic and mPTP analyses but showed limited resolution in distance-based methods (ABGD and ASAP). Overall, the concordance among the majority of species delimitation methods and loci provides robust evidence supporting the recognition of C. veneta sp. nov. as a distinct species (Figure 1).

3.5 Morphometric analysis

Univariate analyses of the 33 morphometric variables from 40 specimens (20 Chromis veneta and 20 Chromis viridis; Table 6) revealed significant interspecific differences in 15 of the measured characters, whereas the remaining variables did not differ significantly between Chromis veneta and C. viridis. Body depth (BD) exhibited the strongest discrimination between species (F = 217.23, P < 0.0001), followed by caudal concavity (CCC; F = 94.20, P < 0.0001), pre pelvic length (PPL; F = 46.05, P < 0.0001), head length (HL; F = 28.35, P < 0.0001), maximum caudal peduncle depth (CPDmax; F = 20.98, P < 0.0001), predorsal length (PDL; F = 16.55, P < 0.001), eye diameter (ED; F = 16.50, P < 0.001), caudal peduncle length (CPL; F = 10.68, P = 0.0023), second dorsal-fin spine length (2DS; F = 10.48, P = 0.0025), snout length (SnL; F = 9.23, P = 0.0043), head depth (HD; F = 8.44, P = 0.0061), upper jaw length (UJL; F = 7.75, P = 0.0083), lower dorsal-fin ray length (LDFR; F = 5.60, P = 0.023), fourth dorsal-fin spine length (4DS; F = 5.26, P = 0.027), and last dorsal spine length (LDS; F = 4.53, P = 0.040). In contrast, soft dorsal fin base (SDB), first and third dorsal-fin spine lengths (1DS and 3DS), pre anal length (PAL), first and second anal-fin spine lengths (1AS and 2AS), lower anal-fin ray length (LAR), and anal-fin base length (AFB) showed no significant differences between species (P > 0.05).

Multivariate analysis of variance (MANOVA) based on all 33 morphometric variables demonstrated highly significant overall differences between C. veneta and C. viridis (Wilks’ λ = 0.0406, approximate F = 16.45, P = 2.89 × 10⁻⁷). Equivalent results were obtained using Pillai’s Trace (0.9594), Hotelling–Lawley Trace (23.64), and Roy’s Largest Root (23.64), all of which were highly significant (P < 0.001), confirming a strong multivariate separation between the two taxa.

Principal component analysis (PCA) further supported the morphological differentiation between the proposed new species and C. viridis, resolving two largely distinct clusters with only minimal overlap among a few individuals. (Figure 4). The first principal component (PC1) explained 25.6% of the total morphometric variation and the second principal component (PC2) explained 16.4%, together accounting for 42.0% of the overall variation. PC1 primarily represented interspecific variation in body proportions and was strongly influenced by body depth (14.39% contribution), caudal concavity (11.04%), head length (9.01%), pre dorsal length (8.76%), maximum caudal peduncle depth (8.42%), second dorsal-fin spine length (6.59%), pre pelvic length (6.13%), caudal peduncle length (5.28%), snout length (5.20%), lower dorsal-fin base length (5.05%), eye diameter (4.11%) and upper jaw length (3.81%). In contrast, PC2 was influenced primarily by lower anal-fin ray length (18.13%), first anal-fin spine length (17.88%), second anal-fin spine length (14.20%), anal-fin base length (13.33%), pre anal length (7.02%), soft dorsal base (6.71%) and first dorsal-fin spine length (6.03%), reflecting variation largely within species rather than between species. Although slight overlap occurred near the group boundaries, the overall separation of the two species was distinct and consistent with the results of the MANOVA and LDA analyses. Linear discriminant analysis (LDA) produced a single discriminant function that clearly separated the two species (Figure 5). The largest standardized coefficients on LD1 were associated with body depth (1.352), second anal-fin spine length (−1.134), soft dorsal fin base (0.312), snout length (0.280), caudal concavity (−0.267), eye diameter (0.232), fourth dorsal-fin spine length (0.175), second dorsal-fin spine length (0.170), first anal-fin spine length (0.141), and caudal peduncle length (−0.162), indicating that these variables contributed most strongly to discrimination between species. Leave-one-out cross-validation correctly classified 38 of the 40 specimens (95%), with only a single specimen of C. veneta and one specimen of C. viridis being misclassified. The high classification success, together with the pronounced separation observed in PCA, demonstrates that the two taxa are morphometrically well differentiated despite limited overlap among a few individuals.

Figure 4

Figure 5

Permutational multivariate analysis of variance (PERMANOVA) based on Euclidean distances detected significant differences in overall morphometric composition between C. veneta and C. viridis (pseudo-F = 12.98, R² = 0.255, P = 0.0001; 9,999 permutations), indicating that species identity explained approximately 25.5% of the total multivariate variation. However, the test for homogeneity of multivariate dispersions (PERMDISP) was also significant (F = 16.16, P = 0.0003), demonstrating unequal within-species multivariate dispersion, with C. viridis exhibiting greater morphological variability than C. veneta. Because PERMANOVA can be influenced by heterogeneity of dispersion, these results were interpreted cautiously. Nevertheless, the concordant evidence from MANOVA, PCA, LDA, and the univariate analyses consistently supports the recognition of Chromis veneta as a morphometrically distinct species.

3.6 Description of a new species, Chromis veneta sp. nov.

English name: Laccadive blue green chromis.

LSIDurn:lsid:zoobank.org:act:07FE1AD8-CB04-47FE-9731-C1FB3F68E035.

3.6.1 Holotype

NBFGR/POMCVEN (52 mm SL), Agatti, Lakshadweep, India. 10°49’23.2”N and 72°10’26.5”E, associated with Acropora corals, 2m, Lift net, Rekha M. U., 26 October 2024 (Figures 6–10, Table 5).

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

3.6.2 Paratypes

NBFGR/POMCVEN.1 (49.3 mm SL), NBFGR/POMCVEN.2 (48.35 mm SL), NBFGR/POMCVEN.3 (45.06 mm SL), NBFGR/POMCVEN.5 (54.63 mm SL), Near to Vallinokkam, Gulf of Mannar, India. 9°2’35.9”N 78°40’59.2”E, associated with Acropora corals, 1.5 m, Rekha M. U and T. T. Ajith Kumar, 24 January 2025. NBFGR/POMCVEN.4 (50.61 mm SL), NBFGR/POMCVEN.6 (43.03 mm SL), Agatti, Lakshadweep, India.10°49’23.2”N and 72°10’26.5”E, associated with Acropora corals, 2m, Lift net, Rekha M. U., 26 October 2024. NBFGR/POMCVEN.7 (47.89 mm SL), NBFGR/POMCVEN.8 (62.4 mm SL), NBFGR/POMCVEN.9 (64.51 mm SL), Agatti, Lakshadweep, India.10°52’23.1”N and 72°11’47.4”E, associated with Acropora corals, 2m, Lift net, Rekha M. U., 15 December 2024.

3.6.3 Diagnosis

Chromis veneta sp. nov. is distinguished from its congeners by the following combination of characters: dorsal-fin rays XII, 9–10 (usually 10); anal-fin rays II, 9–10 (usually 10); pectoral-fin rays 16–17 (usually 17); pelvic-fin rays I, 5; principal caudal-fin rays 8 + 7; spiniform caudal rays 3; tubed lateral-line scales 15–16; gill rakers 9–10 + 23–26 (usually 9 + 24), total 33–35 (usually 33); body depth 2.27–2.56 in SL; caudal-peduncle depth 0.63–0.77 in head length; and caudal-peduncle length 0.51–0.73 in head length. In life, the body is bright aquamarine-green with a subtle iridescent blue-green sheen. Each scale on the upper and middle flanks is narrowly edged with yellow-green, forming a fine reticulate pattern that becomes more conspicuous under illumination. The ventral surface is pale silvery white. A diffuse black patch is present at the upper base of the pectoral fin and is more conspicuous in live specimens. A fine, continuous blue-green stripe extends from the tip of the snout to the anterior margin of the orbit. Reproductively active individuals further exhibit distinct yellowish to amber pigmentation along the distal margins of the dorsal and anal fins.

3.6.4 Description

Dorsal-fin rays XII,10 (rarely XII,9); anal-fin rays II,10 (rarely II,9). All dorsal- and anal-fin soft rays branched, except the first anal-fin soft ray, which is occasionally unbranched. Pectoral-fin rays 17 (rarely 16), uppermost 2 and lowermost rays unbranched; pelvic-fin rays I,5. principal caudal-fin rays 8 + 7 (=15); procurrent caudal-fin rays 5 dorsally and 5 ventrally, the anterior three rays spiniform and the posterior two segmented and unbranched. Tubed lateral-line scales 15–16; posterior midlateral scales with a pore or deep pit 8–10 (usually 10); scale rows above lateral line to origin of dorsal fin 2; scale rows below lateral line to origin of anal fin 8. Gill rakers 9 to 10 + 23 to 26 (total 32–35, usually 9 + 24 = 33). Branchiostegal rays 6; supraneural (predorsal) bones 3; vertebrae 11 + 15.

Morphometric measurements are expressed as percentages of standard length (% SL) in Table 5. In the text, measurements are reported either as proportions of standard length (SL) or as proportions of head length (HL), as appropriate. Body moderately elongate and compressed, body depth 2.47 (2.27–2.56) in SL; body width 2.90 (2.68–3.43) in body depth. Head length 3.59 (3.32–3.74) in SL Dorsal profile of head gently convex from snout to anterior margin of eye, nearly straight to slightly concave across the orbital region, and gently convex posteriorly across the interorbital region and nape. Snout shorter than orbital diameter, its length 4.07 (3.34–4.81) in head length; orbit diameter 3.37 (3.16–3.78) in head length. Interorbital region convex, its width 3.08 (2.85–3.39) in head length. Caudal peduncle moderately slender, its depth 2.16 (2.01–2.32) in head length and its length 1.67 (1.43–2.00) in head length.

Mouth terminal, small, and oblique, with upper jaw forming an angle of approximately 40° to the longitudinal axis of the body. Posterior margin of maxilla extending slightly beyond a vertical through anterior edge of pupil; upper-jaw length 2.38 (2.19–2.64) in head length. Lower jaw bearing a single row of robust, conical, moderately recurved teeth. Six or seven enlarged, caniniform teeth present anteriorly on dentary, accompanied by very small accessory teeth on the inner side, forming a conspicuous symphyseal series, with median teeth largest. Teeth posterior to this series progressively decreases in size laterally while retaining a conical, recurved form. Dentary teeth well-spaced and directed inward; accessory teeth and inner tooth rows absent. Upper jaw with a single row of numerous small, slender, conical teeth along margin of premaxilla. Premaxillary teeth closely set, subequal in size, acute, and weakly recurved, lacking enlarged caniniform teeth. Teeth directed ventrally and slightly posteriorly; inner tooth rows absent. Tongue with a triangular tip.

Gill rakers long and slender, the longest rakers on lower limb near angle of first gill arch approximately four-fifths length of longest gill filaments. Nostrils lacking a fleshy rim, posterior portion slightly elevated; nostril situated at level of middle of pupil and slightly less than one-third of distance from tip of snout to base of upper lip.

Opercle terminating posteriorly in a flattened spine, its tip relatively blunt and largely concealed by a large scale. Preopercular margin weakly crenulate, with posterior edge extending dorsally to level of upper margin of pupil. Suborbital with a free ventral margin extending nearly to a vertical through posterior edge of pupil.

Scales finely ctenoid. Anterior portion of lateral line terminating below posterior part of spinous dorsal fin, between the twelfth and thirteenth dorsal-fin spines. Head scaled except on lips, tip of snout, and a narrow naked region extending from orbit to anterior snout margin enclosing the nostrils. A well-developed scaly sheath present along base of anal fin. Anal-fin membranes bearing a single row of scales that decrease progressively in size distally. Caudal fin scaled basally, with small scales extending slightly beyond two-thirds of fin length toward posterior margin. A median scaly process projects posteriorly between bases of pelvic fins, its length approximately one-half that of pelvic spine; axillary scale above pelvic-fin base also about one-half length of pelvic spine.

Origin of dorsal fin situated above fourth lateral-line scale; predorsal length 2.93 (2.71–3.15) in SL. Base length of spinous dorsal fin 2.80 (2.70–2.97) in SL; base length of soft-rayed dorsal fin 7.82 (6.77–9.67) in SL. Dorsal-fin spines progressively increase in length to third spine and thereafter gradually decrease posteriorly; first dorsal-fin spine 10.49 (9.41–11.91) in SL, second 7.33 (6.63–8.10), third 7.41 (6.89–7.91), fourth 7.78 (7.36–8.33), fifth 8.07 (7.66–8.56), sixth 8.33 (7.86–8.78), and last dorsal-fin spine 9.84 (8.67–11.10) in SL.

Membranes between dorsal-fin spines not incised. Soft-rayed portion of dorsal fin pointed, with fifth soft ray longest, its length 5.08 (4.47–5.78) in SL. First anal-fin spine short, 23.45 (20.37–28.62) in SL; second anal-fin spine longer and more robust, 9.27 (8.43–10.14) in SL. Fifth anal-fin soft ray longest, its length 5.17 (4.79–5.53) in SL.

Caudal fin deeply forked, with pronounced filamentous extensions on both lobes; caudal-fin length 2.77 (2.39–3.38) in SL and caudal concavity 4.44 (3.77–5.19) in SL. Pectoral fin pointed, with first ray longest, its length 4.14 (3.68–4.58) in SL. Pelvic fin elongate; pelvic-fin spine 7.67 (7.25–8.76) in SL, and first pelvic-fin soft ray longest, produced into a filament, its length 3.94 (3.50–4.46) in SL.

Colour when freshly collected (Figure 6). Adults iridescent turquoise to blue-green, with the dorsal region exhibiting intense emerald to neon-green iridescence, gradually fading ventrally to pale silvery white, forming a pronounced dorsoventral colour gradient. Upper half of body with each scale bearing a bright turquoise to green reflective centre narrowly bordered by yellow-green, producing a distinct reticulate to finely tessellated appearance under reflected light; no continuous longitudinal stripe present. Lower flanks and abdomen silvery white with dense iridescent blue and green speckling. Head metallic turquoise-blue with strong reflective sheen; snout and interorbital region slightly deeper blue-green. Opercular region silvery with bluish-green reflections, lacking any distinct ocellus or blotch. Iris silvery golden with metallic reflective margins; pupil black. Dorsal fin translucent to hyaline with bluish-green suffusion along fin rays; distal margin pale yellow to amber, more distinct posteriorly. Caudal fin deeply forked, translucent bluish with elongate lobes and bright cyan-blue margins. Anal fin bright yellow to orangish, hyaline distally. Individuals smaller than approximately 50 mm SL lack the distinctive yellow or amber pigmentation on the distal portions of the dorsal and anal fins, which develops only in larger individuals. Pelvic fins translucent with faint bluish tinge. Pectoral fins hyaline and transparent. A faint dusky blotch is occasionally present at the upper base of the pectoral fin and is more conspicuous in live specimens. A narrow iridescent blue-green line extending from snout tip to anterior margin of orbit.

During reproductive period (Figure 7), the breeding males are more slender and predominantly bright yellow to yellow-green. Dorsum and upper flanks with intense chartreuse to emerald-green iridescence, becoming increasingly metallic posteriorly. Anterior half of body relatively uniform yellow-green, with scales weakly outlined. Posterior half of body and caudal peduncle with each scale bearing a vivid emerald to turquoise-green reflective centre bordered by broad blackish margins, forming a conspicuous reticulate to lattice-like pattern. Ventral region and lower flanks paler yellowish green to whitish. Head uniformly yellow-green; cheek and opercular regions with metallic turquoise reflections. Snout, lips, and lower jaw pale yellow. Operculum lacking distinct ocellus or dark blotch. Iris dark brown to black with narrow metallic golden outer ring; pupil black. A distinct dark bluish-black blotch present at upper base of pectoral fin. Dorsal fin translucent yellow, fin rays yellow to amber, distal margin narrowly dusky. Caudal fin deeply forked; basal and median portions dusky to blackish, lobes with strong iridescent turquoise to blue-green reflections, particularly along rays and posterior margins. Distinct bright yellow to orange pigmentation is present along the anal fin and pelvic fin. Breeding females become increasingly greenish yellow in coloration. A distinct feature is the presence of bright yellow to orange pigmentation along the distal margins of the anal fin and elongate pelvic fin filaments.

Colouration (postmortem, Figure 8). Body uniformly greyish to bluish grey, markedly duller than in life; dorsal region slightly darker, grey to slate-grey; flanks grey with faint bluish tinge; ventral region pale grey to whitish. Iridescent blue-green hues and distinct scale-centre spotting present in life largely lost, remaining only as weakly discernible, diffuse darker centres under reflected light. Head concolorous with body; snout and interorbital region slightly darker. Opercular region greyish, lacking distinct markings or opercular blotch. Eye dark; iris dull grey to blackish. Dorsal fin translucent greyish; spines and rays dusky, distal margin faintly yellowish, more evident on posterior soft portion. Anal fin hyaline to pale grey with weak yellowish tinge distally. Caudal fin pale grey, semi-translucent, without distinct markings. Pelvic fins pale grey to slightly yellowish hyaline. Pectoral fins hyaline.

Colour in preservative (Figure 9). The specimen appears generally dark grey to blackish, with the dorsal region distinctly darker than the ventral surface. The upper head, nape, and dorsum are uniformly dusky to blackish-grey, gradually fading to pale silvery-grey on the lower flanks and abdomen. The opercular and suborbital regions are silvery with prominent metallic blue to bluish-violet iridescent patches. A faint a diffuse black patc is present at the upper base of pectoral fin, becoming less distinct in preservative. The snout and lower jaw are pale brownish to flesh-coloured. The fins are mostly dusky to translucent; the dorsal fin membranes are greyish-hyaline with darker fin rays and faint dusky pigmentation distally. The caudal fin is dusky grey, becoming paler towards the distal portions, with darker pigmentation concentrated along the fin rays and outer margins. The anal fin is dark grey to brownish-grey, with darker fin rays and slightly translucent margins. The pectoral and pelvic fins are hyaline to pale grey, with pigmentation restricted mainly to the fin-ray bases. The caudal peduncle is uniformly dark grey, concolorous with the posterior body. Eye silvery white in preservative, with pupil obscured by preservation.

3.6.5 Distribution

Currently, Chromis veneta sp. nov. is confirmed from coral reef habitats of the Lakshadweep Archipelago in the southeastern Arabian Sea and the Gulf of Mannar along the southeast coast of India, based on the type series and additional examined material. Published molecular data further indicate that this lineage also occurs in the western and central Indo-Pacific, including Madagascar, Indonesia (Jakarta Raya and Komodo), the Great Barrier Reef of Australia, and Fiji, where specimens were previously identified as C. viridis but are herein assignable to C. veneta sp. nov.

3.6.6 Etymology

The specific epithet, veneta is derived from the Latin venetus, meaning “sea-blue” or “azure,” referring to the characteristic of blue green to turquoise coloration exhibited by the species in live condition. The name is treated as an adjective in agreement with the generic name Chromis.

3.6.7 Remarks

It was previously recognized as one of the cryptic lineages within the Chromis viridis complex (Liu et al., 2019; Messmer et al., 2012). Its close external resemblance to C. viridis likely contributed to its historical misidentification and omission from previous taxonomic accounts. The present study demonstrates that the species is diagnosable by a combination of morphometric, meristic, coloration, and otolith characters, supported by concordant mitochondrial and nuclear genetic divergence. The species inhabits shallow Acropora-dominated reef habitats and occurs in sympatry with C. viridis in the Lakshadweep Archipelago (Figure 11).

Figure 11

3.7 Comparisons

Chromis veneta sp. nov. was compared with two congeners, Chromis viridis and C. atripectoralis, using both literature data and specimens examined in the present study (Tables 6, 7). The new species differs from Chromis viridis in several diagnostic morphological characters, including body pigmentation pattern, dentition and caudal spinule morphology.

Chromis veneta sp. nov. is readily distinguished from Chromis viridis by its more vivid and structured colouration pattern. New species possesses a bright turquoise-blue ground colour overlaid with strong yellow-green iridescence, producing a conspicuous speckled and reticulated pattern extending across most of the flanks; the scales are strongly outlined with yellow to yellow-green reflective centres, imparting a pronounced metallic net-like appearance (Figure 6). In contrast, Chromis viridis exhibits a relatively uniform emerald to blue-green dorsum that gradually fades ventrally to pale bluish white, with yellow spots embedded within the blue-green pigmentation mainly restricted to the upper and middle regions of the body and lacking a distinct reticulate pattern (Figure 12).

Figure 12

Chromis viridis differs from Chromis veneta sp. nov. in several external morphological characters. In Chromis viridis, the distribution of melanophores is interrupted by distinct yellow spots lacking melanophores, producing a mottled pigmentation pattern (Figure 13a). In contrast, new species exhibits a more homogeneous distribution of melanophores, with yellow pigmentation primarily bordering the scales and lacking the distinct melanophore-free spots observed in C. viridis (Figure 13b).

Figure 13

Dentition also differs markedly between the two taxa. The lower jaw of C. viridis bears 6–8 enlarged, caniniform teeth accompanied by numerous smaller teeth arranged in multiple inner rows (Figure 13c). In contrast, Chromis veneta sp. nov. possesses a single row of 6–7 enlarged, caniniform teeth on the lower jaw, accompanied by comparatively fewer inner rows of smaller teeth or accessory teeth (Figure 13d).

In Chromis viridis, the pectoral-fin base blotch comprises elongate, diffusely distributed melanophores that extend posteriorly along the fin base and adjacent scale rows (Figure 13e). In C. veneta sp. nov., melanophores are smaller, more rounded, and densely aggregated at the anterior region above the pectoral fin base (Figure 13f).

Additionally, the caudal peduncular spines exhibit distinct interspecific differences. In Chromis viridis, the spines are comparatively longer, more triangular in profile, and only weakly recurved (Figure 13g), whereas those of Chromis veneta sp. nov. are relatively short, robust, and strongly recurved posteriorly (Figure 13h).

Sagittal otolith morphology also differs between the two species. The sagitta of Chromis viridis is relatively broad and robust, with a more irregular outline and comparatively pronounced development of the rostral and antirostral regions. The dorsal and posterior margins are distinctly crenulated, and the medial surface exhibits conspicuous sculpturing, including prominent radial ridges and grooves (Figure 13i). In contrast, the sagitta of C. veneta sp. nov. is comparatively narrower and more elongate, with a relatively smoother outline and less pronounced development of the rostrum and antirostrum. The dorsal and posterior margins are less strongly crenulated, and the medial surface exhibits comparatively weaker radial sculpturing, resulting in a relatively more slender and less ornamented otolith profile (Figure 13j).

Compared with C. viridis, the new species differs markedly in postmortem colouration, particularly in its darker overall body pigmentation and more pigmented fins (Figures 8, 14). Chromis viridis exhibits a predominantly silvery-grey body with weak dorsoventral countershading, faint melanophore speckling on the dorsum, and largely hyaline fins with only slight dusky pigmentation on the dorsal and caudal fin rays. In contrast, the new species possesses a darker dusky bluish-grey to dark grey body coloration, with dense melanophore concentration on the dorsal region extending onto the upper flanks and caudal peduncle. The flanks of the new species additionally show faint bluish iridescence, absent or indistinct in C. viridis. The fins of the new species are also more heavily pigmented: the dorsal fin bears a faint yellowish distal tint, while the caudal fin is dark grey to blackish with distinctly prolonged filamentous upper and lower lobes. In C. viridis, the caudal fin lacks filamentous extensions and remains lightly dusky to hyaline. Anal and pelvic fins in the new species are translucent to pale yellowish with lightly pigmented distal margins, whereas those of C. viridis are translucent whitish and weakly pigmented. Both species lack conspicuous bars, stripes, ocelli, or lateral markings in preserved condition.

Figure 14

Pectoral-fin ray counts in C. veneta sp. nov. range from 16–17, whereas C. viridis exhibits a broader range of 15–17, including the lower count of 15, which is absent in C. veneta. The most diagnostic differences, however, occur in the gill-raker counts. Chromis veneta sp. nov. differs from C. viridis in having higher numbers of upper gill-rakers (9–10 vs. 8–9) and lower gill-rakers (23–26 vs. 21–23), resulting in a higher total gill-raker count (32–35 vs. 29–32). C. viridis. In contrast, vertebral counts are identical in both taxa, with 11 precaudal and 15 caudal vertebrae recorded in both C. veneta sp. nov. and C. viridis (Figures 10, 15; Tables 5, 8).

Figure 15

The new species also differs from C. atripectoralis in having a shallower body depth (39.1–44.0% SL vs. 45.5–47.5% SL), shorter head length (25.3–30.1% SL vs. 29.3–31.4% SL), smaller eye diameter (7.8–8.7% SL vs. 9.3–11.3% SL), shorter snout length (5.8–7.8% SL vs. 7.3–8.7% SL), and a narrower interorbital width (8.2–9.9% SL vs. 9.0–10.5% SL). In addition, C. veneta is distinguished by its remarkably shorter pelvic fins (12.7–13.8% SL vs. 23.0–28.1% SL) and comparatively shorter caudal peduncle length (13.7–20.2% SL vs. 17.0–22.2% SL). Chromis veneta exhibits lower pectoral-fin ray counts (16–17), whereas C. atripectoralis shows distinctly higher counts ranging from 16–19, most commonly 17–18. This higher pectoral-ray count is one of the principal diagnostic characters separating C. atripectoralis from related blue green chromis species. Gill-raker counts also provide strong separation between the species. Chromis veneta possesses comparatively higher upper gill-raker counts (9–10) than C. atripectoralis (7–10, usually 8–9). Lower gill-raker counts in C. veneta range from 23–26, distinctly exceeding the counts observed in C. atripectoralis (19–22). Consequently, total gill-raker counts are substantially higher in C. veneta (32–35) compared with C. atripectoralis (28–31) (Table 6).

4 Discussion

4.1 Resolving cryptic diversity in the Chromis viridis species complex

Previous molecular studies identified a deeply divergent lineage within Chromis viridis based solely on mitochondrial/nuclear sequence divergence and interpreted it as a putative cryptic species (Froukh and Kochzius, 2008; Messmer et al., 2012; Hubert et al., 2017; Liu et al., 2019). However, the lineage was not formally described because its molecular distinctiveness was not supported by an integrative taxonomic assessment incorporating morphological evidence. Consequently, this lineage remained an unnamed component of the C. viridis complex despite its repeated recovery in molecular phylogenetic studies.

Our integrative analyses resolve this long-standing taxonomic uncertainty. Multilocus phylogenetic inference (COI, Cytb, D-loop, RAG1, and RAG2), species delimitation analyses, quantitative morphometrics, meristics, and colour pattern consistently identify this lineage as an independently evolving and morphologically distinct species. It is readily distinguished from C. viridis by a significantly more slender body, a narrower caudal peduncle, higher gill-raker counts, differences in dentition and caudal fin spinule morphology, distinctive adult colouration, and a characteristic otolith morphology.

These findings demonstrate that the lineage does not represent a cryptic species in the strict sense, as it is diagnosable using multiple independent morphological and molecular characters. Rather, its recognition was delayed by the morphological conservatism of the blue-green chromis complex and the lack of comprehensive integrative taxonomic studies. Accordingly, the lineage previously recognised as “Chromis viridis Clade B” by Liu et al. (2019) is herein formally described as Chromis veneta sp. nov. Its recognition increases the number of valid species in the blue-green chromis complex from two to three and suggests that additional overlooked diversity may remain within the widespread C. viridis complex across the Indo-Pacific.

4.2 Sympatry and morphological conservatism obscure species boundaries

The discovery of C. veneta is a remarkable consideration as the species have immense commercial importance in the marine ornamental fish trade. Blue green chromis are among the most heavily traded coral reef fishes globally and are routinely harvested in large numbers for aquarium export (Wabnitz et al., 2003; Rhyne et al., 2012). Despite decades of extensive collection, international trade, and captive maintenance, this lineage remained taxonomically overlooked. This suggests that strong morphological conservatism, overlapping coloration patterns, and reliance on superficial external characters have collectively masked species-level diversity within the complex. Notably, photographs identified as C. viridis across aquarium websites, hobbyist forums, and online databases frequently exhibit conspicuous variation in body coloration, scale reflectance, dorsal pigmentation, and fin coloration. Such variation may reflect the historical aggregation of multiple cryptic taxa under a single commercial identity.

Unlike many recently described reef fishes that occupy geographically isolated regions, C. veneta occurs sympatrically with C. viridis across parts of its distribution, including the Lakshadweep Archipelago and the Gulf of Mannar. Such coexistence, coupled with their remarkably similar external appearance, made species discrimination extremely difficult. Currently, Chromis veneta sp. nov. is confirmed from coral reef habitats of the Lakshadweep Archipelago in the southeastern Arabian Sea and the Gulf of Mannar along the southeast coast of India, based on the type series and additional examined material. Beyond its confirmed Indian distribution, genetically verified records are available from Komodo Island (Indonesia), Fiji, and Lizard Island on the northern Great Barrier Reef, Australia (Messmer et al., 2012; Liu et al., 2019). In previous molecular studies, specimens from these localities were recovered as Clade B within the Chromis viridis species complex, and the occurrence of this lineage in sympatry with C. viridis sensu stricto was also documented (Liu et al., 2019). Furthermore, BLAST searches of COI sequences identified three additional GenBank accessions from Jakarta Raya, Indonesia (FJ583146–FJ583148), originally deposited as Chromis caerulea, and one sequence from Toliara, Madagascar (OL410087), deposited as Chromis sp., that showed high sequence similarity to C. veneta sp. nov. These records indicate that the species is likely more widely distributed across the Indo-West Pacific than currently recognized, although much of its range has probably remained undetected because of historical misidentification as morphologically similar congeners.

The inclusion of Indian populations in the present study considerably extends the confirmed distribution of this lineage. Nevertheless, based on currently verified records, C. veneta sp. nov. exhibits a discontinuous distribution across the Indo-West Pacific, contrasting with the broad, continuous Indo-Pacific range historically attributed to the Chromis viridis species complex. Additional sampling throughout the region will be required to determine the full extent of its distribution. This disjunct distribution may reflect historical biogeographic processes, incomplete sampling across intervening regions, or the low relative abundance of C. veneta in areas where it co-occurs with C. viridis, as suggested by Liu et al. (2019). The recognition of this distribution pattern has likely been further hindered by the long-standing assumption that blue-green damselfishes represented a single, morphologically variable species throughout the Indo-Pacific. Consequently, geographic variation was generally interpreted as intraspecific polymorphism rather than evidence of distinct evolutionary lineages. Moreover, the widespread reliance on external colouration for species identification obscured lineage diversity because members of the C. viridis complex exhibit remarkably conserved colour patterns despite substantial genetic divergence.

The commercial importance of the blue-green chromis has likely contributed to perpetuating this taxonomic oversight. Despite the collection of millions of individuals for the marine ornamental trade over several decades, specimens have routinely been identified as C. viridis without genetic verification. Consequently, the recognition of C. veneta demonstrates how commercial exploitation, sympatry, morphological conservatism, and historical taxonomic assumptions have collectively masked species diversity in coral reef fishes.

4.3 Historical misidentification of Chromis caerulea and Chromis viridis

The recognition of Chromis veneta sp. nov. may also have been influenced by the historical assignment of sequences belonging to this lineage to C. viridis, C. caerulea, or unidentified Chromis taxa in public databases. Although the extent of its impact cannot be quantified, such instances are consistent with broader challenges in DNA reference libraries, where morphology-based identifications and incomplete voucher documentation can lead to inaccuracies in public sequence databases (Collins and Cruickshank, 2013; Weigt et al., 2012). These inaccuracies may, in turn, affect phylogenetic inference, influence estimates of species distributions, and obscure underlying lineage diversity (Bemis et al., 2023; Keck et al., 2023). Similar issues have been reported for marine fishes, where cryptic diversity and unresolved taxonomy can limit the reliability of molecular reference databases (Victor, 2015; Weigt et al., 2012).

In the blue-green damselfishes, these problems are further compounded by a complex nomenclatural history. The long-standing taxonomic ambiguity surrounding the application of the name Chromis caerulea was comprehensively reviewed by Randall et al. (1985), who traced its origin to the interpretation of Jordan and Seale (1906). Following the examination of the type specimens by Léon Vaillant, Jordan and Seale (1906) regarded Heliases caeruleus, H. lepisurus, and H. frenatus as conspecific and consequently adopted Chromis caeruleus as the valid name for the Blue-green Damselfish. Randall et al. (1985) demonstrated that this interpretation was inconsistent with the International Code of Zoological Nomenclature because Bleeker (1877), acting as the First Reviser, had already selected Heliases lepisurus as the valid name among these simultaneously published taxa. To resolve this long-standing nomenclatural instability, Randall et al. (1987) submitted a formal application to the International Commission on Zoological Nomenclature (ICZN) proposing the conservation of the name Chromis viridis and the suppression of the misapplied name C. caerulea. The proposal was accepted in ICZN Opinion 1563 (ICZN, 1989), which suppressed the specific name caeruleus Cuvier, 1830, as published in the binomen Heliases caeruleus, for the purposes of the Principle of Priority but not those of the Principle of Homonymy. The name was placed on the Official Index of Rejected and Invalid Specific Names in Zoology, whereas the specific name ternatensis Bleeker, 1856 was placed on the Official List of Specific Names in Zoology. This ruling stabilized the nomenclature of the blue-green damselfishes and established Chromis viridis as the valid name for the Blue-green Damselfish.

Randall et al. (1985) also clarified the identity of Chromis viridis by rediscovering Cuvier’s (1830) original description of Pomacentrus viridis, which was based on two unpublished colour illustrations, one by Ehrenberg from Massawa in the Red Sea and the other by Mertens from Guam. Because Cuvier explicitly attributed the species to Ehrenberg, Randall et al. (1985) designated the Red Sea illustration, which depicts an approximately 8 cm long specimen, as the lectotype of Pomacentrus viridis. Although the Guam illustration could represent either C. viridis or C. atripectoralis, the Red Sea specimen can be unequivocally assigned to C. viridis because C. atripectoralis does not occur in the Red Sea (Randall et al., 1985). The lectotype illustration depicts a relatively deep-bodied, laterally compressed fish that conforms to the morphology of C. viridis or C. atripectoralis. By contrast, Chromis veneta sp. nov. consistently exhibits a more elongate and streamlined body, as demonstrated by the present morphometric analyses, and adults (> 6.5 cm TL) possess characteristic yellow to amber pigmentation along the fin margins, a feature absent from the lectotype of C. viridis. These morphological differences, together with the multilocus genetic evidence presented here, further support the recognition of C. veneta as a distinct species.

Several COI sequences corresponding to Chromis veneta sp. nov. have previously been deposited in GenBank under the name Chromis caerulea, reflecting historical misidentifications resulting from an unresolved species complex. As C. caerulea is a suppressed junior synonym of C. ternatensis, its application to this lineage is nomenclaturally invalid. The integrative evidence presented in this study unequivocally demonstrates that this lineage is distinct from both C. viridis and C. ternatensis, warranting recognition as a separate species. Consequently, the designation of this lineage as a cryptic species is no longer appropriate, as C. veneta sp. nov. is diagnosable by a consistent suite of morphological, meristic, colour pattern, and molecular evidence.

The recognition of Chromis veneta sp. nov. resolves a long-standing taxonomic ambiguity within the C. viridis species complex and provides a robust framework for re-evaluating existing sequence records in public databases. Correct assignment of these records will improve the accuracy of molecular identifications, refine our understanding of species distributions, and strengthen future systematic, phylogenetic, and evolutionary studies of Chromis.

4.4 Morphological differentiation and taxonomic recognition of Chromis veneta sp. nov.

Species recognition within Chromis has traditionally relied on subtle but consistent combinations of external morphological characters because meristic characters are highly conserved throughout the genus. Most species possess XII–XIII dorsal-fin spines, II anal-fin spines, and overlapping numbers of dorsal-, anal- and pectoral-fin rays, rendering fin-ray counts alone insufficient for species diagnosis (Allen, 1991; Tang et al., 2021). Consequently, modern taxonomic revisions have increasingly emphasized body colouration, body proportions, caudal-fin morphology, gill-raker counts and osteological characters as the principal diagnostic features for distinguishing closely related species (Allen and Allen, 2021; Tang et al., 2021).

Among these characters, body depth represents one of the most informative morphometric traits and exhibits considerable interspecific variation within Chromis. Species range from exceptionally deep-bodied forms, such as Chromis bowesi (body depth 1.5–1.6 in SL; Arango et al., 2019), C. caudalis (1.65–1.85 in SL; Randall, 1988), C. ternatensis (1.83–2.02 in SL; Randall et al., 1981) and C. alpha (approximately 1.8–2.0 in SL; Randall, 1988), to comparatively slender-bodied species such as C. viridis (2.04 – 2.21 in SL; Randall et al., 1981) C. atripectoralis (2.02–2.16 in SL; Randall et al., 1981), C. gunting (2.1–2.2 in SL; Arango et al., 2019) and C. norfolkensis (2.0 – 2.4 in SL; Allen and Allen, 2021). This wide range of body proportions demonstrates that body depth has diversified substantially among evolutionary lineages and remains one of the most informative characters for species diagnosis within the genus.

The present study provides the first comprehensive morphological evidence supporting the recognition of the ‘Clade B’ lineage of Chromis viridis as a distinct species. Chromis veneta sp. nov. is characterised by a distinctly more slender body (body depth 2.27–2.56 in SL; 39.11–43.99% SL), placing it among the more slender-bodied species of the genus Chromis. In contrast, C. viridis possesses a substantially deeper body (2.04 – 2.21 in SL), resulting in a more robust body profile (Randall et al., 1981). The new species is further distinguished by its narrower caudal peduncle, more deeply forked caudal fin with elongate lobe tips, bright yellow to orange distal margins on the dorsal and anal fins in adults, higher gill raker count, and differences in the pattern of dentition, and sagittal otolith morphology.

Colour pattern has long been recognised as one of the most informative external characters for distinguishing closely related species of Chromis, particularly within morphologically conservative groups in which meristic characters exhibit considerable overlap (Welander and Schultz, 1951; Randall et al., 1981; Allen, 1991). The blue-green chromis complex is a classic example of such morphological conservatism. Until recently, it comprised only two recognised species, Chromis viridis and C. atripectoralis, despite their widespread Indo-Pacific distributions and extensive overlap in meristic characters and overall body morphology. Based on sympatric populations, Welander and Schultz (1951) demonstrated that these two species are most reliably distinguished by the pigmentation of the pectoral-fin axil. Chromis viridis has only a faint, dusky pigmentation in the pectoral-fin axil, formed by scattered melanophores and largely confined to the dorsal portion of the axil. In contrast, C. atripectoralis has a conspicuous, solid black axillary blotch with a sharply defined outer margin. Chromis veneta sp. nov. lacks the distinct black axillary blotch characteristic of C. atripectoralis and more closely resembles C. viridis in having only diffuse axillary pigmentation. However, C. veneta is readily distinguished from both species by its distinctive body colouration. Although all three species exhibit a superficially similar blue-green colouration, C. veneta sp. nov. possesses a bright turquoise-blue body with strong yellow-green iridescence and a conspicuous reticulated pattern produced by yellow to yellow-green pigmentation outlining the scales, particularly over the upper and middle regions of the body. In contrast, both C. viridis and C. atripectoralis exhibit yellow pigmentation as discrete spots embedded within the blue-green background, lacking the characteristic reticulated appearance of C. veneta sp. nov.

These qualitative differences are corroborated by morphometric analyses. Fifteen of the thirty-three size-adjusted morphometric variables differed significantly between C. veneta and C. viridis, with body depth providing the greatest discriminatory power, followed by caudal concavity, head length, pre pelvic length and caudal peduncle depth. Multivariate morphometric analyses consistently supported the distinctiveness of the two species. MANOVA revealed highly significant overall differences in body shape, PCA recovered two largely discrete morphospaces with minimal overlap, and linear discriminant analysis correctly assigned 95% of specimens to their respective species. Although C. viridis exhibited significantly greater within-group multivariate dispersion (PERMDISP), PERMANOVA demonstrated that the species differed significantly in multivariate centroid position. The concordance among univariate morphometric comparisons and multiple multivariate approaches provides strong evidence that the observed morphological differences represent species-level divergence rather than geographic variation within a single species.

Gill-raker counts have traditionally been considered of limited diagnostic value within the blue-green chromis complex because of the substantial overlap between Chromis viridis and C. atripectoralis. In the original description of Chromis atripectoralis, Welander and Schultz (1951) distinguished the species from the taxon then recognized as Chromis caerulea (now C. viridis) by its generally lower total gill-raker counts (28–31 vs. 29–33). However, the substantial overlap in gill-raker counts between the two species indicates that this character alone is insufficient for reliable species identification and should be interpreted in conjunction with other morphological and molecular evidence. In contrast, Chromis veneta sp. nov. exhibits a consistently higher total gill-raker count (32 – 35) than both C. viridis and C. atripectoralis, providing an additional diagnostic character for distinguishing species within the blue-green chromis complex. Although gill-raker counts alone should not be used in isolation for species delimitation, their concordance with the distinctive body shape, colour pattern, morphometric differentiation, and multilocus molecular evidence substantially strengthens the diagnosis of C. veneta sp. nov.

Within Chromis, C. veneta sp. nov. is hypothesised to belong to the subgenus Hoplochromis based on the combination of morphological characters that define this lineage. Fowler and Bean (1928) first recognised the distinctiveness of the chromis assemblage by proposing the subgenus Hoplochromis, diagnosing it primarily by the presence of six short outward-flaring conical teeth on the anterior margin of the lower jaw. Subsequent revisions refined the diagnosis by recognising three exposed spiniform procurrent rays on both the upper and lower margins of the caudal-fin base as a principal synapomorphy of the group, together with a conservative fin formula and shared external morphology (Randall et al., 1981; Tang et al., 2021). Chromis veneta sp. nov. possesses both the six outward-flaring conical teeth and three exposed spiniform procurrent caudal-fin rays, supporting the hypothesis that it belongs to the Hoplochromis lineage alongside C. atripectoralis, C. flavaxilla, C. ternatensis and C. viridis. Nevertheless, confirmation of its phylogenetic placement within the subgenus will require a comprehensive genus-level analysis including a broader representation of Chromis species.

The morphological evidence is fully congruent with the multilocus phylogenetic reconstruction and species-delimitation analyses presented herein. This concordance is consistent with the principles of integrative taxonomy, whereby independent lines of evidence provide reciprocal support for species hypotheses (Dayrat, 2005; Padial et al., 2010; Tang et al., 2021). Collectively, the molecular, morphometric and meristic evidence demonstrates that the ‘Clade B’ lineage of Chromis viridis (Liu et al., 2019) represents an independently evolving and morphologically distinct species, herein recognised as Chromis veneta sp. nov.

4.5 Species delimitation and evolutionary divergence within the Chromis viridis complex

In the present study, multilocus phylogenetic analyses based on mitochondrial (COI, cytb, and D-loop) and nuclear (RAG1 and RAG2) markers consistently recovered Chromis veneta sp. nov. as a distinct lineage within the C. viridis complex. All mitochondrial markers showed strong differentiation of C. veneta from its congeners, indicating the presence of lineage-specific genetic variation. The higher resolution provided by mitochondrial loci likely reflects their relatively faster evolutionary rates and smaller effective population size, which facilitate more rapid lineage sorting during recent divergence events (Avise, 2000; Hebert et al., 2003). The concordant recovery of C. veneta across multiple mitochondrial markers reduces the possibility that the observed genetic distinctiveness represents a locus-specific pattern.

Although the nuclear RAG1 and RAG2 marker exhibited comparatively lower phylogenetic resolution than mitochondrial loci, it independently recovered C. veneta sp. nov. as a separate evolutionary lineage. The reduced resolution of nuclear markers and variation in the placement of some congeners among nuclear gene trees are consistent with expected patterns of gene-tree heterogeneity. Nuclear loci generally evolve more slowly and possess larger effective population sizes than mitochondrial loci, resulting in longer persistence of ancestral polymorphisms and slower lineage sorting, which may produce discordance between individual gene trees and the underlying species tree (Degnan and Rosenberg, 2009; Edwards, 2009). Therefore, the differences observed among individual gene trees are interpreted as a consequence of locus-specific evolutionary histories rather than conflicting evidence regarding species boundaries.

Similar mitochondrial–nuclear discordance has been reported in other Chromis species complexes. For example, limited resolution and signatures of incomplete lineage sorting in nuclear markers have been documented among closely related taxa such as C. dimidiata, C. fieldi, and C. margaritifer, whereas mitochondrial markers provided stronger differentiation among lineages (He et al., 2019). These observations suggest that differences in resolution among loci are common in recently diverged reef fish lineages and highlight the importance of integrating multiple independent genetic markers for accurate species delimitation.

The recognition of C. veneta sp. nov. is further supported by the concordance among multiple species-delimitation approaches. Distance-based methods (ABGD and ASAP), tree-based approaches (mPTP), and the multispecies coalescent framework (BPP) consistently identified C. veneta as an independent lineage. Although individual delimitation methods differ in their assumptions and sensitivity to evolutionary processes, agreement among multiple approaches provides strong evidence that the observed genetic discontinuity represents a true species boundary rather than an artefact of a single analytical method (Puillandre et al., 2012; Yang and Rannala, 2010).

The discovery of C. veneta highlights the underestimated diversity concealed within widespread coral reef fishes and demonstrates the importance of integrative approaches for resolving species boundaries in morphologically conservative groups. Recognition of this lineage provides a revised evolutionary framework for the C. viridis complex and emphasizes the need for taxonomically validated reference sequences to improve future biodiversity assessments, ecological studies, and molecular identifications.

4.6 Evolutionary and conservation implications

Genetic diversity represents the evolutionary foundation of biodiversity by maintaining adaptive potential, ecological resilience, and the capacity for long-term evolutionary persistence (Frankham et al., 2002; Hughes et al., 2008). Consequently, failure to recognize distinct evolutionary lineages may result in an underestimation of biodiversity and the loss of unique evolutionary histories represented by those lineages (Faith, 1992; Moritz, 2002). The concordance among multilocus phylogenetic analyses, species-delimitation approaches, and diagnostic morphological characters in the present study highlights the effectiveness of integrative taxonomy for resolving species boundaries in morphologically conservative coral reef fishes (Dayrat, 2005; Pante et al., 2015). Recognition of Chromis veneta sp. nov. therefore refines species limits within the C. viridis complex and contributes to a more accurate understanding of reef fish diversity, evolutionary relationships, and Indo-Pacific biogeographic patterns.

The pronounced genetic divergence observed among lineages suggests long-term evolutionary independence rather than recent population subdivision. In the present study, Chromis veneta sp. nov. was consistently recovered as a distinct lineage across mitochondrial markers (COI, Cyt b, and D-loop), and its independent recovery in the nuclear RAG1 and RAG2 phylogeny provides additional support for its evolutionary distinctiveness within the C. viridis complex. Although RAG2 showed comparatively lower resolution and differences in the placement of some congeners, the consistent separation of C. veneta sp. nov. indicates that the observed mitochondrial divergence is not restricted to a single locus but reflects broader lineage differentiation. Such multilocus divergence may arise and persist even in reef fishes with high dispersal potential, as pelagic larval development does not necessarily result in unrestricted gene flow. Increasing evidence demonstrates that effective connectivity among reef fish populations is frequently limited by larval behaviour, oceanographic circulation patterns, habitat discontinuity, and local retention processes (Jones et al., 2005; Selkoe and Toonen, 2011; Saenz-Agudelo et al., 2011). These mechanisms can promote genetic subdivision and facilitate the persistence of independently evolving lineages despite potentially high dispersal capacity (Barber et al., 2002; Luiz et al., 2012). Therefore, historical restrictions in gene flow and localized evolutionary processes may have contributed to the persistence and diversification of lineages within the C. viridis complex.

The observed phylogeographic structure is consistent with historical isolation and reduced connectivity among Indo-Pacific reef habitats. Repeated Pleistocene sea-level fluctuations substantially altered shallow-water habitats, intermittently fragmenting and reconnecting reef systems and creating opportunities for lineage divergence (Voris, 2000). Such historical biogeographic processes, together with dispersal limitation and oceanographic barriers, have contributed to diversification patterns observed in coral reef fishes across the Indo-Pacific region (Carpenter et al., 2011; Crandall et al., 2012). Subsequent secondary contact among previously isolated lineages, particularly within the Indo-Malay-Philippine biodiversity region, may have facilitated the coexistence of closely related taxa while allowing the persistence of morphological similarity under comparable reef environments (Rocha et al., 2005; Gaither and Rocha, 2013). The sympatric occurrence of Chromis viridis and C. veneta sp. nov. is consistent with this scenario, where historical isolation followed by secondary contact may contribute to lineage coexistence within the C. viridis complex. Similar patterns of cryptic diversification associated with ecological conservatism have been documented across several coral reef fish groups (Leray et al., 2010; Hubert et al., 2017).

Despite limited morphological differentiation, cryptic lineages often represent distinct evolutionary histories with differences in demographic trajectories, genetic structure, and ecological associations (Tornabene et al., 2015; Underwood et al., 2019). Ecological differentiation and habitat specialization may further contribute to reproductive isolation among closely related reef fish lineages occurring in sympatry, facilitating the persistence of independently evolving lineages (Puebla et al., 2014; Underwood et al., 2019). Together, these patterns suggest that diversification within the Chromis viridis complex has likely resulted from the interaction of historical isolation, restricted gene flow, and contemporary ecological processes.

Recognition of independently evolving lineages has important implications for biodiversity assessment, conservation planning, and sustainable management of marine resources. Failure to distinguish multiple evolutionary lineages within a widespread nominal species can obscure geographically restricted taxa, underestimate biodiversity, and complicate assessments of species distributions and conservation status (Bickford et al., 2007). Such overlooked diversity is particularly relevant in marine systems, where morphologically similar lineages may remain undetected until molecular approaches reveal previously unrecognized evolutionary units (Bickford et al., 2007; Bowen, 2016).

The recognition of Chromis veneta sp. nov. has important implications for marine ornamental trade management and aquaculture practices. Historically, genetically distinct but morphologically similar lineages may have been grouped under a single species name, potentially obscuring differences in their geographic distribution, population structure, abundance, and vulnerability to exploitation. This issue is particularly relevant for ornamental aquaculture, where C. viridis has recently emerged as a promising candidate for captive propagation, with studies documenting its captive breeding protocols, spawning behaviour, and early ontogenetic development (Rekha et al., 2024). However, reliance on morphology-based identification alone may result in the inclusion of genetically distinct lineages within broodstock populations, complicating the management of lineage-specific genetic diversity and the interpretation of breeding performance. Accurate genetic validation of broodstock identity is therefore essential to ensure reliable aquaculture production, maintain genetic integrity, and avoid unintended mixing of independently evolving lineages. Such lineage mixing may complicate broodstock management, population monitoring, and the preservation of lineage-specific genetic diversity, particularly when genetically distinct taxa are maintained under a single species name. Similar concerns regarding overlooked genetic diversity and the importance of accurate lineage identification have been highlighted in other marine fish conservation and aquaculture contexts (Richards and Hobbs, 2015). Therefore, recognition of C. veneta sp. nov. emphasizes the need for genetically validated species identification in both wild-capture fisheries and ornamental aquaculture, where accurate taxonomic assignment is fundamental for sustainable management, effective breeding programmes, and conservation of evolutionary lineages.

5 Conclusion

Our integrative taxonomic assessment demonstrates that the lineage previously interpreted as cryptic diversity within Chromis viridis is neither cryptic nor conspecific with C. viridis but represents a distinct evolutionary lineage that is formally described herein as a new species. The failure to recognize this species for decades resulted from its superficial resemblance to C. viridis, sympatric occurrence, historical nomenclatural confusion surrounding Chromis caerulea, and the lack of comprehensive voucher-based taxonomic studies. By integrating evidence from five genetic markers (COI, Cytb, D-loop, RAG1 and RAG2) with detailed morphometric, meristic, and colour pattern analyses, this study resolves the long-standing taxonomic uncertainty within the Chromis viridis complex. The formal recognition of the new species, Chromis veneta sp. nov. refines our understanding of species diversity in Indo-Pacific coral reef fishes and underscores the importance of integrative taxonomy in uncovering overlooked biodiversity and informing conservation strategies.

References

• 1

  AllenG. R. (1991). Damselfishes of the World (Melle: Mergus Publishers).

  • Google Scholar
  • View reference in article

• 2

  AllenG. R.AllenM. G. (2021). Two new species of Chromis (Teleostei: Pomacentridae) from northwestern Australia and the southwestern Pacific Ocean, previously part of C. fumea (Tanaka 1917). J. Ocean Sci. Found.38, 78–103. doi: 10.5281/zenodo.5601970

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  AndersonM. J. (2001). A new method for non-parametric multivariate analysis of variance. Austral Ecol.26, 32–46. doi: 10.1111/j.1442-9993.2001.01070.pp.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  AndersonM. J. (2006). Distance-based tests for homogeneity of multivariate dispersions. Biometrics62, 245–253. doi: 10.1111/j.1541-0420.2005.00440.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  ArangoB. G.PinheiroH. T.RochaC.GreeneB. D.PyleR. L.CopusJ. M.et al. (2019). Three new species of Chromis (Teleostei, Pomacentridae) from mesophotic coral ecosystems of the Philippines. ZooKeys835, 1–16. doi: 10.3897/zookeys.835.27528

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  AviseJ. C. (2000). Phylogeography: The History and Formation of Species (Cambridge, MA: Harvard University Press).

  • Google Scholar
  • View reference in article

• 7

  BarberP. H.PalumbiS. R.ErdmannM. V.MoosaM. K. (2002). Sharp genetic breaks among populations of Haptosquilla pulchella (Stomatopoda) indicate limits to larval transport: patterns, causes, and consequences. Mol. Ecol.11, 659–674. doi: 10.1098/rspb.2002.2026

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  BemisK. E.GirardM. G.SantosM. D.CarpenterK. E.DeedsJ. R.PitassyD. E.et al. (2023). Biodiversity of Philippine marine fishes: A DNA barcode reference library based on voucher specimens. Sci. Data10, 411. doi: 10.1038/s41597-023-02306-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  Ben-TzviO.AbelsonA.PolakO.KiflawiM. (2008a). Habitat selection and the colonization of new territories by Chromis viridis. J. Fish Biol.73, 1005–1018. doi: 10.1111/j.1095-8649.2008.02003.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  Ben-TzviO.KiflawiM.GainesS. D.Al-ZibdahM.SheehyM. S.ParadisG. L.et al. (2008b). Tracking recruitment pathways of Chromis viridis in the Gulf of Aqaba using otolith chemistry. Mar. Ecol. Prog. Ser.359, 229–238. doi: 10.3354/meps07343

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  BickfordD.LohmanD. J.SodhiN. S.NgP. K. L.MeierR.WinkerK.et al. (2007). Cryptic species as a window on diversity and conservation. Trends Ecol. Evol.22, 148–155. doi: 10.1016/j.tree.2006.11.004

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  BowenB. W. (2016). The three domains of conservation genetics: case histories from Hawaiian waters. J. Hered.107, 309–317. doi: 10.1093/jhered/esw018

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  CarpenterK. E.BarberP. H.CrandallE. D.Ablan-LagmanM. C. A.AmbariyantoMahardikaG. N.et al. (2011). Comparative phylogeography of the Coral Triangle and implications for marine management. J. Mar. Sci.2011, 396982. doi: 10.1155/2011/396982

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  CollinsR. A.CruickshankR. H. (2013). The seven deadly sins of DNA barcoding. Mol. Ecol. Resour.13, 969–975. doi: 10.1111/1755-0998.12046

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  CooperW. J.SmithL. L.WestneatM. W. (2009). Exploring the radiation of a diverse reef fish family (Labridae) using phylogenetics based on mitochondrial and nuclear DNA. Mol. Phylogenet. Evol.52, 120–131. doi: 10.1016/j.ympev.2008.12.010

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  CrandallE. D.SbroccoE. J.DeBoerT. S.BarberP. H.CarpenterK. E. (2012). Expansion dating: calibrating molecular clocks in marine species from expansions onto the Sunda Shelf following the last glacial maximum. Mol. Biol. Evol.29, 707–719. doi: 10.1093/molbev/msr227

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  DayratB. (2005). Towards integrative taxonomy. Biol. J. Linn. Soc85, 407–417. doi: 10.1111/j.1095-8312.2005.00503.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18

  DegnanJ. H.RosenbergN. A. (2009). Gene tree discordance, phylogenetic inference and the multispecies coalescent. Trends Ecol. Evol.24, 332–340. doi: 10.1016/j.tree.2009.01.009

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  EdwardsS. V. (2009). Is a new and general theory of molecular systematics emerging? Evolution, 63 (1), 1–19.

  • Google Scholar
  • View reference in article

• 20

  ElliottN. G.HaskardK.KoslowJ. A. (1995). Morphometric analysis of orange roughy (Hoplostethus atlanticus) off the continental slope of southern Australia. J. Fish Biol.46, 202–220. doi: 10.1111/j.1095-8649.1995.tb05962.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  FaithD. P. (1992). Conservation evaluation and phylogenetic diversity. Biol. Conserv.61, 1–10. doi: 10.1016/0006-3207(92)91201-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  FauvelotC.BernardiG.PlanesS. (2003). Reductions in the mitochondrial DNA diversity of coral reef fish provide evidence of population bottlenecks resulting from Holocene sea-level change. Evolution57, 1571–1583. doi: 10.1111/j.0014-3820.2003.tb00365.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  FisherR. A. (1936). The use of multiple measurements in taxonomic problems. Ann. Eugenics7, 179–188. doi: 10.1111/j.1469-1809.1936.tb02137.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  FowlerH. W.BeanB. A. (1928). Contributions to the biology of the Philippine Archipelago and adjacent regions. The fishes of the families Pomacentridae, Labridae, and Callyodontidae, collected by the United States Bureau of Fisheries Steamer “Albatross” chiefly in Philippine seas and adjacent waters. Bull. U.S. Natl. Mus.100, 1–525. Available online at: https://cir.nii.ac.jp/crid/1573668924858200192.

  • Google Scholar
  • View reference in article

• 25

  FrankhamR.BriscoeD. A.BallouJ. D. (2002). Introduction to Conservation Genetics (Cambridge, UK: Cambridge University Press). doi: 10.1017/CBO9780511808999

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  FrickeR.EschmeyerW. N.Van der LaanR. (2026). Eschmeyer's Catalog of Fishes: Genera, Species, References. Available online at: http://researcharchive.calacademy.org/research/ichthyology/catalog/fishcatmain.asp (Accessed 2026).

  • Google Scholar
  • View reference in article

• 27

  FroukhT.KochziusM. (2008). Species boundaries and evolutionary lineages in the blue-green damselfishes Chromis viridis and Chromis atripectoralis (Pomacentridae). J. Fish Biol.72, 451–457. doi: 10.1111/j.1095-8649.2007.01746.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  GaitherM. R.RochaL. A. (2013). Origins of species richness in the Indo‐Malay‐Philippine biodiversity hotspot: evidence for the centre of overlap hypothesis. J. Biogeogr.40, 1638–1648. doi: 10.1111/jbi.12126

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  HallT. A. (1999). BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser.41, 95–98. doi: 10.1017/CBO9780511819049.025

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  HeS.RobitzchV.HobbsJ. P. A.TraversM. J.Lozano-CortésD.BerumenM. L.et al. (2019). An examination of introgression and incomplete lineage sorting among three closely related species of chocolate-dipped damselfish (genus: Chromis). Ecol. Evol.9, 5468–5478. doi: 10.1002/ece3.5142

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  HebertP. D. N.RatnasinghamS.deWaardJ. R. (2003). Barcoding animal life: Cytochrome c oxidase subunit 1 divergences among closely related species. Proc. R. Soc. B.270, S96–S99. doi: 10.1098/rsbl.2003.0025

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  HubertN.DettaiA.PruvostP.CruaudC.KulbickiM.MyersR. F.et al. (2017). Geography and life history traits account for the accumulation of cryptic diversity among Indo-West Pacific coral reef fishes. Mar. Ecol. Prog. Ser.583, 179–193. doi: 10.3354/meps12316

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  HughesA. R.InouyeB. D.JohnsonM. T.UnderwoodN.VellendM. (2008). Ecological consequences of genetic diversity. Ecol. Lett.11, 609–613. doi: 10.1111/j.1461-0248.2008.01179.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  International Commission on Zoological Nomenclature (ICZN) (1989). Opinion 1563. Bull. Zool. Nomencl.46, 213–214. Avaible online at: https://biostor.org/reference/5126.

  • Google Scholar
  • View reference in article

• 35

  IrwinD. M.KocherT. D.WilsonA. C. (1991). Evolution of the cytochrome b gene of mammals. J. Mol. Evol.32, 128–144. doi: 10.1007/BF02515385

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  IshikawaK.WuH.MitaraiS.GeninA. (2025). Use of videos to measure dynamic body acceleration as a proxy for metabolic costs in coral reef damselfish (Chromis viridis). J. Exp. Biol.228, jeb249717. doi: 10.1242/jeb.249717

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  IvanovaN. V.ZemlakT. S.HannerR. H.HebertP. D. N. (2007). Universal primer cocktails for fish DNA barcoding. Mol. Ecol. Notes7, 544–548. doi: 10.1111/j.1471-8286.2007.01748.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  JohnsonR. A.WichernD. W. (2007). Applied multivariate statistical analysis. 6th ed.Upper Saddle River, NJ: Pearson Prentice Hall.

  • Google Scholar
  • View reference in article

• 39

  JolliffeI. T. (2002). Principal Component Analysis. 2nd Edn (New York, USA: Springer).

  • Google Scholar
  • View reference in article

• 40

  JonesG. P.PlanesS.ThorroldS. R. (2005). Coral reef fish larvae settle close to home. Curr. Biol.15, 1314–1318. doi: 10.1016/j.cub.2005.06.061

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  JordanD. S.SealeA. (1906). The Fishes of Samoa: Description of the Species Found in the Archipelago, With a Provisional Check-List of the Fishes of Oceania Vol. 25 ( Bulletin of the United States Bureau of Fisheries), 173-455.

  • Google Scholar
  • View reference in article

• 42

  KalyaanamoorthyS.MinhB. Q.WongT. K.von HaeselerA.JermiinL. S. (2017). ModelFinder: Fast model selection for accurate phylogenetic estimates. Nat. Methods14, 587–589. doi: 10.1038/nmeth.4285

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  KapliP.LutteroppS.ZhangJ.KobertK.PavlidisP.StamatakisA.et al. (2017). Multi-rate Poisson tree processes for single-locus species delimitation under maximum likelihood and Markov chain Monte Carlo. Bioinformatics33, 1630–1638. doi: 10.1093/bioinformatics/btx025

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  KeckF.CoutonM.AltermattF. (2023). Navigating the seven challenges of taxonomic reference databases in metabarcoding analyses. Mol. Ecol. Resour.23, 742–755. doi: 10.1111/1755-0998.13746

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  KumarS.StecherG.SuleskiM.SanderfordM.SharmaS.TamuraK. (2024). MEGA12: Molecular Evolutionary Genetics Analysis version 12 for adaptive and green computing. Mol. Biol. Evol.41, msae263. doi: 10.1093/molbev/msae263

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  LecchiniD.NakamuraY.GrignonJ.TsuchiyaM. (2006). Evidence of density-independent mortality in a settling coral reef damselfish, Chromis viridis. Ichthyol. Res.53, 298–300. doi: 10.1007/s10228-006-0340-8

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47

  LeeW. J.ConroyJ.HowellW. H.KocherT. D. (1995). Structure and evolution of teleost mitochondrial control regions. J. Mol. Evol.41, 54–66. doi: 10.1007/BF00174041

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  LerayM.BeldadeR.HolbrookS. J.SchmittR. J.PlanesS.BernardiG. (2010). Allopatric divergence and speciation in coral reef fish: the three-spot dascyllus, Dascyllus trimaculatus, species complex. Evolution, 64 (5), 1218–1230.

  • Google Scholar

• 49

  LetunicI.BorkP. (2024). Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res.52, W78–W82. doi: 10.1093/nar/gkae268

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 50

  LiuS. Y. V.TuanmuM. N.RachmawatiR.MahardikaG. N.BarberP. H. (2019). Integrating phylogeographic and ecological niche approaches to delimitating cryptic lineages in the blue–green damselfish (Chromis viridis). PeerJ7, e7384. doi: 10.7717/peerj.7384

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51

  LuizO. J.MadinJ. S.RobertsonD. R.RochaL. A.WirtzP.FloeterS. R. (2012). Ecological traits influencing range expansion across large oceanic dispersal barriers: insights from tropical Atlantic reef fishes. Proc. R. Soc B.279, 1033–1040. doi: 10.1098/rspb.2011.1525

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52

  MessmerV.JonesG. P.MundayP. L.PlanesS. (2012). Concordance between genetic and species diversity in coral reef fishes across the Pacific Ocean biodiversity gradient. Evolution66, 3902–3917. doi: 10.1111/j.1558-5646.2012.01725.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 53

  MoritzC. (2002). Strategies to protect biological diversity and the evolutionary processes that sustain it. Syst. Biol.51, 238–254. doi: 10.1080/10635150252899752

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 54

  MussiM.HaimbergerT. J.HawryshynC. W. (2005). Behavioural discrimination of polarized light in the damselfish Chromis viridis (family Pomacentridae). J. Exp. Biol.208, 3037–3046. doi: 10.1242/jeb.01750

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 55

  MinhB. Q.SchmidtH. A.Von ChernomorO.SchrempfD.WoodhamsM. D.Von HaeselerA.LanfearR. (2020). Q-TREE 2: New models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol.37, 1530–1534. doi: 10.1093/molbev/msaa015

  • CrossRef
  • Google Scholar

• 56

  NylanderJ. A. A. (2004). MrModeltest v2.3. Program distributed by the author (Uppsala, Sweden: Evolutionary Biology Centre, Uppsala University).

  • Google Scholar

• 57

  PadialJ. M.MirallesA.De la RivaI.VencesM. (2010). The integrative future of taxonomy. Front. Zool.7, 16. doi: 10.1186/1742-9994-7-16

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 58

  PalumbiS. R. (1996). “ Nucleic acids II: The polymerase chain reaction,” in Molecular Systematics. Eds. HillisD. M.MoritzC.MableB. K. (Sunderland, MA: Sinauer Associates), 205–247. doi: 10.1093/sysbio/45.4.607

  • CrossRef
  • Google Scholar

• 59

  PanteE.SchoelinckC.PuillandreN. (2015). From integrative taxonomy to species description: one step beyond. Syst. Biol.64, 152–160. doi: 10.1093/sysbio/syu083

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 60

  PueblaO.BerminghamE.McMillanW. O. (2014). Genomic atolls of differentiation in coral reef fishes (Hypoplectrus spp., Serranidae). Molecular Ecology, 23 (21), 5291–5303.

  • Google Scholar

• 61

  PuillandreN.BrouilletS.AchazG. (2021). ASAP: Assemble species by automatic partitioning. Mol. Ecol. Resour.21, 609–620. doi: 10.1111/1755-0998.13281

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 62

  PuillandreN.LambertA.BrouilletS.AchazG. (2012). ABGD, Automatic Barcode Gap Discovery for primary species delimitation. Mol. Ecol.21, 1864–1877. doi: 10.1111/j.1365-294X.2011.05239.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 63

  PyleR. L.EarleJ. L.GreeneB. D. (2008). Five new species of the damselfish genus Chromis (Perciformes: Labroidei: Pomacentridae) from deep coral reefs in the tropical western Pacific. Zootaxa1671, 3–31. doi: 10.11646/zootaxa.1671.1.2

  • CrossRef
  • Google Scholar

• 64

  QuenouilleB.BerminghamE.PlanesS. (2004). Molecular systematics of the damselfishes (Teleostei: Pomacentridae): Bayesian phylogenetic analyses of mitochondrial and nuclear DNA sequences. Mol. Phylogenet. Evol.31, 66–88. doi: 10.1016/S1055-7903(03)00278-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 65

  R Core Team (2025). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. https://www.R-project.org/.

  • Google Scholar

• 66

  RambautA.DrummondA. J.XieD.BaeleG.SuchardM. A. (2018). Posterior summarization in Bayesian phylogenetics using Tracer 1.7. Syst. Biol.67, 901–904. doi: 10.1093/sysbio/syy032

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 67

  RandallJ. E.BauchotM. L.DesoutterM. (1985). Chromis viridis (Cuvier 1830), the correct name for the Indo-Pacific damselfish, previously known as C. caerulea Cuvier 1830 (Pisces, Pomacentridae). Cybium9, 411–413. doi: 10.26028/cybium/1985-94-010

  • CrossRef
  • Google Scholar

• 68

  RandallJ. E.BauchotM.-L.DesoutterM. (1987). Heliases ternatensis Bleeker 1856 (currently Chromis ternatensis; Osteichthyes, Perciformes): proposed conservation, and adoption of the name Chromis viridis (Cuvier 1830) for the fish commonly called C. caerulea (Cuvier 1830). Bull. Zool. Nomencl.44, 248–250.

  • Google Scholar

• 69

  RandallJ. E.IdaH.MoyerJ. T. (1981). A review of the damselfishes of the genus Chromis from Japan and Taiwan, with description of a new species. Japanese J. Ichthyology28, 203–242. doi: 10.5962/bhl.part.345

  • CrossRef
  • Google Scholar

• 70

  RandallJ. E. (1988). Three new Indo-Pacific damselfishes of the genus Chromis (Pomacentridae). Memoirs of the Museum of Victoria, 49 (1), 73–81. doi: 10.11369/jji1950.28.203

  • CrossRef
  • Google Scholar

• 71

  RandallJ. E. (1994). Two new damselfishes (Perciformes: Pomacentridae) from Arabian waters. Revue Française d’Aquariologie Herpétologie, 21, 39–48.

  • Google Scholar

• 72

  RannalaB.YangZ. (2013). Improved reversible jump algorithms for Bayesian species delimitation. Genetics194, 245–253. doi: 10.1534/genetics.112.149039

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 73

  RekhaM. U.HaslamathbiT.BharathiS.AkashS.ShilpaP. N.KumarT. A.et al. (2024). Understanding the captive breeding, spawning behavior and early ontogeny of Chromis viridis (Cuvier 1830): a systematic study. Aquaculture589, 740966. doi: 10.1016/j.aquaculture.2024.740966

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 74

  RhyneA. L.TlustyM. F.SchofieldP. J.KaufmanL. E. S.MorrisJ. A.BrucknerA. W. (2012). Revealing the appetite of the marine aquarium fish trade: the volume and biodiversity of fish imported into the United States. PloS One7, e35808. doi: 10.1371/journal.pone.0035808

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 75

  RichardsZ. T.HobbsJ. P. A. (2015). Hybridisation on coral reefs and the conservation of evolutionary novelty. Curr. Zool.61, 132–145. doi: 10.1002/ece3.6068

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 76

  RochaL. A.RobertsonD. R.RomanJ.BowenB. W. (2005). Ecological speciation in tropical reef fishes. Proc. R. Soc B. Biol. Sci.272, 573–579. doi: 10.1098/2004.3005

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 77

  RonquistF.TeslenkoM.Van Der MarkP.AyresD. L.DarlingA.HöhnaS.et al. (2012). MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst. Biol.61, 539–542. doi: 10.1093/sysbio/sys029

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 78

  Saenz-AgudeloP.JonesG. P.ThorroldS. R.PlanesS. (2011). Connectivity dominates larval replenishment in a coastal reef fish metapopulation. Proc. R. Soc B.278, 2954–2961. doi: 10.1098/rspb.2010.2780

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 79

  SambrookJ.RussellD. W. (2001). Molecular Cloning: A Laboratory Manual. 3rd ed (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press).

  • Google Scholar

• 80

  SelkoeK. A.ToonenR. J. (2011). Marine connectivity: a new look at pelagic larval duration and genetic metrics of dispersal. Mar. Ecol. Prog. Ser.436, 291–305. doi: 10.3354/meps09238

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 81

  SmithM. F. L.WarburtonK. (1992). Predator shoaling moderates the confusion effect in blue-green Chromis, Chromis viridis. Behav. Ecol. Sociobiol.30, 103–107. doi: 10.3390/ani14172572

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 82

  TangC. H.LeuM. Y.ShaoK.HwangL. Y.ChangW. B. (2014). Short-term effects of thermal stress on the responses of branchial protein quality control and osmoregulation in a reef-associated fish, Chromis viridis. Zool. Stud.53, 21. doi: 10.1186/s40555-014-0021-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 83

  TangK. L.StiassnyM. L. J.MaydenR. L.DeSalleR. (2021). Systematics of damselfishes. Ichthyology Herpetology109, 258–318. doi: 10.1643/i2020105

  • CrossRef
  • Google Scholar

• 84

  TornabeneL.ValdezS.ErdmannM.PezoldF. (2015). Support for a ‘Center of Origin’ in the Coral Triangle: cryptic diversity, recent speciation, and local endemism in a diverse lineage of reef fishes (Gobiidae: Eviota). Molecular Phylogenetics and Evolution, 82, 200–210

  • Google Scholar

• 85

  UnderwoodJ. N.TraversM. J.SnowM.PuotinenM.GouwsG. (2019). Cryptic lineages in the Wolf Cardinalfish living in sympatry on remote coral atolls. Mol. Phylogenet. Evol.132, 183–193. doi: 10.1002/eap.2008

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 86

  VictorB. C. (2015). “ How many coral reef fish species are there? Cryptic diversity and the new molecular taxonomy,” in Ecology of Fishes on Coral Reefs ( Cambridge University Press, Cambridge), 76–87. doi: 10.1017/CBO9781316105412.010

  • CrossRef
  • Google Scholar

• 87

  VorisH. K. (2000). Maps of Pleistocene sea levels in Southeast Asia: shorelines, river systems and time durations. J. Biogeogr.27, 1153–1167. doi: 10.1046/j.1365-2699.2000.00489.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 88

  WabnitzC.TaylorM.GreenE. P.RazakT. (2003). From Ocean to Aquarium: The Global Trade in Marine Ornamental Species. UNEP-WCMC Biodiversity Series No. 17. (Cambridge, UK: UNEP World Conservation Monitoring Centre).

  • Google Scholar

• 89

  WeigtL. A.DriskellA. C.BaldwinC. C.OrmosA. (2012). “ DNA barcoding fishes,” in DNA Barcodes: Methods and Protocols ( Humana Press, Totowa, NJ), 109–126.

  • Google Scholar

• 90

  WelanderA. D.SchultzL. P. (1951). Chromis atripectoralis, a new damselfish from the tropical Pacific, closely related to C. caeruleus, family Pomacentridae. J. Washington Acad. Sci.41, 107–110. Available online at: https://www.jstor.org/stable/24533598.

  • Google Scholar

• 91

  WestneatM. W.AlfaroM. E. (2005). Phylogenetic relationships and evolutionary history of the reef fish family Labridae. Mol. Phylogenet. Evol.36, 370–390. doi: 10.1016/j.ympev.2005.02.001

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 92

  YangZ. (2015). The BPP program for species tree estimation and species delimitation. Curr. Zool.61, 854–865. doi: 10.1093/czoolo/61.5.854

  • CrossRef
  • Google Scholar

• 93

  YangZ.RannalaB. (2010). Bayesian species delimitation using multilocus sequence data. Proc. Natl. Acad. Sci.107, 9264–9269. doi: 10.1073/pnas.0913022107

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 94

  YangZ.RannalaB. (2014). Unguided species delimitation using DNA sequence data from multiple loci. Mol. Biol. Evol.31, 3125–3135. doi: 10.1093/molbev/msu279

  • Pubmed Abstract
  • CrossRef
  • Google Scholar