- Research article
- Open Access
Phylogeny and biogeography of a shallow water fish clade (Teleostei: Blenniiformes)
© Lin and Hastings; licensee BioMed Central Ltd. 2013
Received: 9 July 2013
Accepted: 16 September 2013
Published: 25 September 2013
The Blenniiformes comprises six families, 151 genera and nearly 900 species of small teleost fishes closely associated with coastal benthic habitats. They provide an unparalleled opportunity for studying marine biogeography because they include the globally distributed families Tripterygiidae (triplefin blennies) and Blenniidae (combtooth blennies), the temperate Clinidae (kelp blennies), and three largely Neotropical families (Labrisomidae, Chaenopsidae, and Dactyloscopidae). However, interpretation of these distributional patterns has been hindered by largely unresolved inter-familial relationships and the lack of evidence of monophyly of the Labrisomidae.
We explored the phylogenetic relationships of the Blenniiformes based on one mitochondrial (COI) and four nuclear (TMO-4C4, RAG1, Rhodopsin, and Histone H3) loci for 150 blenniiform species, and representative outgroups (Gobiesocidae, Opistognathidae and Grammatidae). According to the consensus of Bayesian Inference, Maximum Likelihood, and Maximum Parsimony analyses, the monophyly of the Blenniiformes and the Tripterygiidae, Blenniidae, Clinidae, and Dactyloscopidae is supported. The Tripterygiidae is the sister group of all other blennies, and the Blenniidae is the sister group of the remaining blennies. The monophyly of the Labrisomidae is supported with the exclusion of the Cryptotremini and inclusion of Stathmonotus, and we elevate two subgenera of Labrisomus to establish a monophyletic classification within the family. The monophyly of the Chaenopsidae is supported with the exclusion of Stathmonotus (placed in the Stathmonotini) and Neoclinus and Mccoskerichthys (placed in the Neoclinini). The origin of the Blenniiformes was estimated in the present-day IndoPacific region, corresponding to the Tethys Sea approximately 60.3 mya. A largely Neotropical lineage including the Labrisomidae, Chaenopsidae and Dactyloscopidae (node IV) evolved around 37.6 mya when the Neotropics were increasingly separated from the IndoPacific, but well before the closure of the Tethys Sea.
Relationships recovered in this study are similar to those of earlier analyses within the Clinidae and Chaenopsidae, and partially similar within the Blenniidae, but tripterygiid relationships remain poorly resolved. We present the first comprehensive phylogenetic hypothesis for a monophyletic Labrisomidae with five tribes (Labrisomini, Mnierpini, Paraclinini, Stathmonotini and Starksiini). Global distributions of blenny genera included in our analysis support the evolution of a largely Neotropical clade whose closest relatives (clinids and cryptotremines) are temperate in distribution.
The Blenniiformes comprises six families, 151 genera and nearly 900 species  of small teleost fishes closely associated with coastal benthic habitats. Although small, blennies often numerically dominate rocky reef ichthyofaunas [2–4]. Because of their abundance and the ease at which they can be observed, blennies have become convenient models for the study of ecology and behavior . Blennies also provide an unparalleled opportunity for marine biogeography because they include the globally distributed families Tripterygiidae and Blenniidae, the temperate Clinidae, and three families (Labrisomidae, Chaenopsidae, and Dactyloscopidae) largely restricted to the Neotropics . The breakup of the Tethys Sea , formation of the Central America Isthmus , and climate warming during the Pliocene  have been suggested as important historical events shaping the evolution and current distributional patterns of blennies. However, insights from these and other studies have been compromised by the absence of a well-resolved phylogeny for the group.
The history of blenniiform classification has been reviewed by several researchers (e.g. [1, 10–12]). Members of the currently recognized families Tripterygiidae, Blenniidae, Labrisomidae, Clinidae and Chaenopsidae are consistently included, but other families have been added depending on the definition of “true” blennies [11, 13–16]. A widely accepted concept of a monophyletic Blenniiformes as the “tropical” blenny families (= Blenniicae sensu Hubbs, 1952; ) was revisited by Springer  and, based on morphological characters, formalized to include the above five families and the sand stargazers, Dactyloscopidae, as hypothesized earlier by Regan . Springer’s Blennioidei, termed the Blenniiformes by Wiley and Johnson , shares several unique morphological features including presence of a bean-shaped pelvis, a reduced branchial apparatus, proximal pectoral-fin radials longer than wide, unbranched pectoral-fin rays, relatively simple caudal-fin morphology, 0–2 spines and simple segmented rays in the anal-fin, and no neural spine on the first vertebra [11, 18].
Inter-familial relationships of blenniiforms have remained largely unresolved because of conflicting morphological and molecular evidence . It is generally agreed that the Tripterygiidae is the sister group of all other blennies, based on these groups having unbranched dorsal- and pectoral-fin rays (branched in most triplfins), no ctenoid scales as in triplefins, and roofed sensory canal bones (unroofed in triplefins) . It has also been hypothesized that the Blenniidae is the sister group of the remaining blennies based on several features of the dorsal gill-arches and associated muscles . However, a consensus has not been reached regarding relationships among the Clinidae, Chaenopsidae, Dactyloscopidae and Labrisomidae. The monophyly of the first three of these families has been supported by morphological synapomorphies . However, the Labrisomidae includes generalized blennies that do not fall into the other relatively well-defined families and no synapomorphies have been identified for this group  other than one possible reductive character . Labrisomids have long been considered to be closely related to the Clinidae (e.g. ) and included in that family by some authors (e.g. ). In addition, the relationships of the Dactyloscopidae to other blennies, the most recent major lineage to be added to the Blenniiformes , have been evaluated based only on dorsal gill-arch anatomy . Thus, dactyloscopids have been placed in an unresolved polytomy along with the Clinidae, Labrisomidae and Chaenopsidae . Two molecular studies have attempted to resolve the phylogenetic relationships among the blenny families [9, 22]. These provided inconsistent phylogenetic hypotheses, although both questioned the monophyly of the Labrisomidae and the Chaenopsidae.
Recently, the systematics of blennies was thoroughly reviewed, providing convenient reference points: the Labrisomidae, Clinidae, Chaenopsidae, and Dactyloscopidae by Hastings and Springer , the Blenniidae by Hastings and Springer  and the Tripterygiidae by Fricke . In addition, studies on higher-level relationships of fishes based on morphology and multiple genetic markers have suggested a close relationship between blennies and clingfishes (Gobiesocidae) [19, 25–30], jawfishes (Opistognathidae) [31, 32], and basslets (Grammatidae) [31, 33–35] within the recently recognized Ovalentaria . In the present study, we attempt to reconstruct the phylogenetic relationships of the Blenniiformes with significantly broader taxon sampling (150 blenniiform species), substantially more genetic information (one mitochondrial and four nuclear markers), and more strategic outgroup representation than in previous studies.
Family (total genera/ total species)
Lineage (total genera/ sampled genera)
Genera (total species/ sampled species)
Salariini-Salarias group (20/12)
Salariini-Rhabdoblennius group (8/1)
Proportions of variable and parsimony informative characters, best nucleotide substitution model and parameters selected by jModelTest for each data partition
Parsimony informative characters
Best model by jModelTest
Among the 3,562 bp, the alignment comprised 1,627 (45.68%) variable sites, of which 1,392 (39.08%) were parsimony informative. The sequences of the ten partitions provided a range of evolutionary information. The third codon position of all five markers provided significantly more parsimony information (59.09% to 99.47%) than the first and second positions (4.13% to 28.21%). The third codon of the only mitochondrial marker (COI) had the highest proportion of variable characters (100%) and the highest proportion of parsimony informative characters (99.47%). The first and second codon positions of the nuclear marker Histone H3 had the lowest proportion of both variable characters (9.63%) and parsimony informative characters (4.13%).
In general, the relationships among triplefins were poorly resolved in all analyses (Figure 2). Relationships within the Blenniidae were, however, well resolved, with only a few discrepancies among analytical methods (Figure 3). The monophyly of all blenniid genera for which we sampled multiple species was well supported except that Alticus saliens was nested within Andamia, and Atrosalaris fuscus was nested within Salarias (Figure 3). The monophyly of each the four blenniid tribes included in our study (i.e., Parablenniini, Salariini, Nemophini, and Omobranchini) was well supported with the exception of the Salariini in the MP analysis (MP bootstrap value = 66). The two Blenniinae tribes Nemophini and Omobranchini were grouped together and were the sister group of the tribe Parablenniini. This entire clade was the sister group of the tribe Salariini.
Relationships among the remaining four families were well resolved in the shallower nodes, but less so in the deeper nodes (Figure 4). In general the pattern is the inclusion of low-diversity taxa as the sister group of larger monophyletic clades. These low-diversity groups include Calliclinus (Cryptotremini), Auchenionchus + Alloclinus (both currently Cryptotremini), and the “chaenopsids” Neoclinus and Mccoskerichthys. In addition to a monophyletic Clinidae, all three analytical methods recovered two diverse clades with moderate support values. These included the members of the Labrisomidae (node II) other than the Cryptotremini that together formed a monophyletic group with the inclusion of the genus Stathmonotus (currently considered a chaenopsid). The second lineage (node III) included the Dactyloscopidae plus the Chaenopsidae sensu stricto (i.e., excluding Stathmonotus, Neoclinus and Mccoskerichthys).
Within the labrisomid clade (node II; BI PP=74), BI recovered relationships with Dialommus (Mnierpini) as the sister group of two relatively large clades, the Paraclinini + Stathmonotus, and Starksiini + Labrisomini (Figure 4). Within the Paraclinini, the monotypic genus Exerpes was nested within the genus Paraclinus, while within the Starksiini, the monotypic genus Xenomedea was the sister group of a monophyletic Starksia (Figure 4). Within the Labrisomini, the genus Labrisomus was not monophyletic, but divided into two clades. One clade included the subgenus Labrisomus that was the sister group of the monophyletic genus Malacoctenus, while the other clade included all other sampled members of Labrisomus (Figure 4).
In the remaining large clade (node III; BI PP=86), the monophyletic Dactyloscopidae was the sister group of a monophyletic Chaenopsidae sensu stricto (i.e., exclusive of Stathmonotus, Neoclinus and Mccoskerichthys). Within the Chaenopsidae all currently recognized genera were recovered as monophyletic, and the genus Coralliozetus was the sister group of all other chaenopsids. That group included two large clades, one with Emblemariopsis, Protemblemaria and Cirriemblemaria, and another with Acanthemblemaria, Ekemblemaria, Hemiemblemaria, Emblemaria, Chaenopsis and Lucayablennius (Figure 4).
The divergence times of blenniiform lineages were estimated from the combined postburn-in trees and parameter values of the BEAST analyses (Additional file 2: Figure S1). The maximum clade credibility tree estimated from the posterior density was characterized by a large number of nodes supported with significant Bayesian posterior probabilities (data not shown) and the mean of the posterior density of the likelihood score was −78678.5184 (95% highest posterior densities [HPD]: -78748.29 to −78607.9508).
Families and inter-family relationships of the Blenniiformes
Although the monophyly of the Blenniiformes sensu Wiley and Johnson, 2009  (= Blennioidei sensu Springer, 1993; ) has been questioned by some (e.g. ), our phylogenetic analysis found strong support for its monophyly in agreement with other recent molecular analyses (e.g. [32, 35]). Blenniiform monophyly is also supported by seven morphological character complexes [1, 11, 18]. This lineage includes nearly 900 species allocated among six families and is the sister group of the Gobiesocidae (Figure 1), a relationship supported by six morphological characters, one of which is unique to this lineage .
Our study supports the monophyly of four of the six blenniiform families as currently construed. The mono-phyly of the Tripterygiidae, Blenniidae, Clinidae and Dactyloscopidae are also well-supported by morphological synapomorphies [1, 11, 23]. The monophyly of the Chaenopsidae (sensu Hastings and Springer, 1994 ) and the Labrisomidae are, however, not supported in our analysis (see below).
Within the Blenniiformes, our study indicates that the Tripterygiidae is the sister group to the remaining blennies (Figure 1). This hypothesis was proposed by Springer  based on details of the fin rays, scales and cephalic sensory system, and subsequently supported by Springer and Orrell  based on details of the branchial musculature. Our analysis also supports the hypothesis of Springer and Orrell  based on several features of the dorsal gill-arches and associated muscles that the Blenniidae is the sister group of the remaining four families of blennies (Figure 1). That clade includes the Clinidae, Labrisomidae, Chaenopsidae and Dactylosco-pidae and is referred herein as the “clinioids” (node I; Figures 1 and 4). Relationships within the clinioids are complicated and not entirely consistent with the current nomenclature, as reported by earlier studies based on morphological characters , allozyme data , and mitochondrial 12S rRNA sequence data . In our analysis, a single species of the labrisomid tribe Cryptotremini (Calliclinus geniguttatus) is the sister group of all other clinioids, and the monophyletic family Clinidae is the sister group of the remaining clinioids (Figure 4). Our analysis provided partial resolution of deep relationships within the latter lineage, but with short branch lengths to their most recent common ancestors (Figure 4), implying frequent divergence events (Additional file 2: Figure S1) consistent with a rapid radiation . Bayesian Inference and Maximum Likelihood analyses support another two cryptotremins (Alloclinus holderi + Auchenionchus microcirrhis) as the sister group of the remaining species, and two “chaenopsids” (Neoclinus blanchardi + Mccoskerichthys sandae) as the sister group of all others (unresolved in ML) (Figure 4). These low diversity branches are followed by two speciose lineages (node IV), one that includes most labrisomid species (node II; Figure 4), and a second, newly identified clade that includes the Chaenopsidae sensu stricto and the Dactyloscopidae (node III; Figure 4).
The Tripterygiidae, comprising 32 genera and 164 species, can be readily distinguished from other blenniiforms by their three-part dorsal fin which lacks the “last” dorsal-fin spine [11, 24]. Two subfamilies and eight tribes of triplefins have recently been proposed based on morphological evidence [24, 39], but their relationships remain unclear. Recent molecular studies on the phylogenetic relationships of triplefins are restricted to regional studies with limited taxon sampling (e.g. [40–42]) and the same is partially true of our study. We included representatives of four tribes, eight genera and 19 species from the subfamily Tripterygiinae (Table 1 and Additional file 1: Table S1) with an emphasis on eastern Pacific species. Although the monophyly of triplefins is strongly supported by our data, there is a general pattern of low node support values within the family (Figure 2). Our poor resolution of relationships within the Tripterygiidae is likely a consequence of inadequate taxon sampling (19 of 164 species), incomplete data sampling (i.e. missing data for some markers), the presence of numerous insertions and deletions observed in the sequences of TMO-4C4 and RAG1 which complicates phylogenetic analyses , and the possible rapid diversification of the group . Clearly much broader taxon sampling and further study will be needed to resolve relationships within the triplefins.
We sampled 20 genera and 48 species, representing four of the six tribes and both subfamilies Blenniinae and Salariinae (Table 1 and Additional file 1: Table S1). Two low diversity tribes, the Blenniini with four species and the monotypic Phenablenniini, were not included in our analysis. The monophyly of the four tribes included in our study is supported (Figure 3), as is the monophyly of the Blenniinae. However, the monophyly of the Salariinae (sensu Bath, 2001 ) is not supported. Instead, we found the Parablenniini to be the sister group of the Blenniinae rather than the Salariini (Figure 6b). However, because our study included few genera of the Parablenniini (2 of 14) and did not include a member of the tribe Blenniini, additional taxon sampling is necessary to more fully resolve the relationships of the major lineages of blenniids.
The Nemophini genera Meiacanthus, Plagiotremus, Petroscirtus, and Xiphasia form a monophyletic clade that is sister to the Omobranchini genus Omobranchus (Figure 3). Within the Nemophini, this analysis provides an alternative phylogeny to the provisional hypothesis proposed by Smith-Vaniz  in which Meicanthus and Petroscirtes together were the sister group of Plagiotremus, Xiphasia, and Aspidontis. In our analysis, Meiacanthus and Plagiotremus form a strongly supported clade that is the sister group of Petroscirtus and Xiphasia (Figure 3). Relationships of the three species of Plagiotremus included in our study are consistent with those proposed by Smith-Vaniz (, Fig 81).
Only two out of the fourteen Parablenniini genera (Hypsoblennius and Parablennius) are included in our study and these form a monophyletic group (Figure 3). The six Hypsoblennius species show incongruent relationships from those suggested by Bath  in several respects.
The monophyly of the Clinidae, comprising 26 genera and 85 species, is supported by the presence of cycloid scales with radii in all fields and a cordlike ligament extending from the ceratohyal to the dentary [11, 14, 53]. Three tribes are recognized within this family [14, 53] and a hypothesized relationship of the Myxodini as the sister group of the Clinini + Ophiclinini is based on the latter two sharing the reproductive pattern of internal fertilization with males possessing an intromittent organ . Our study includes representatives of the Clinini and Myxodini and the monophyly of each is well supported (Figure 4). Species relationships within the Myxodini genus Gibbonsia (Figure 4) are congruent with the previously hypothesized relationships based on 40 allozyme loci , but conflict with a more recent study with expanded taxon sampling based on the same allozyme loci  and 12S rRNA data . Within the Clinini, the Australian genera Heteroclinus and Cristiceps form a well-supported sister group to the South African genera Clinus, Muraenoclinus, Blennophis, and Pavoclinus (Figure 4).
The monophyly of the Labrisomidae has long been questioned because of the lack of any supporting morphological characters [11, 19] or molecular evidence [9, 22]. Also, relationships among the five included tribes remain unclear . Our study includes representatives of all five hypothesized tribes and 10 of the 14 genera (Table 1 and Additional file 1: Table S1). The only genera not included are the Eastern Pacific deepwater (> 20 m depth) genus Cryptotrema with two species (Cryptotremini; [14, 55]), the recently described monotypic Cottoclinus (Mnierpini; ), and two rare and poorly known genera of uncertain relationships, Nemaclinus with one species  and Haptoclinus with two species [21, 58]. The present study thus provides the most thorough investigation to date of the phylogenetic relationships of members of this family.
In our analysis, a monophyletic Labrisomidae (node II; Figure 4) is recovered with the exclusion of the Crypto-tremini and the inclusion of Stathmonotus. In our BI analysis, Stathmonotus is the sister group of the Para-clinini, and that clade is the sister group of a clade including the Starksiini and Labrisomini (Figure 4). However, relationships among the major lineages of labrisomids have strong nodal support only from the BI analysis.
The tribe Cryptotremini is the only exclusively temperate group historically allocated to the Labrisomidae while all others are primarily tropical. This tribe was first described by Hubbs  who included the northeastern Pacific genera Alloclinus and Cryptotrema. The Cryptotremini was later expanded with the addition of two southeastern Pacific genera (Auchenionchus and Calliclinus), although their inclusion was based on the plesiomorphic condition of branched caudal-fin rays . The caudal-fin rays are unbranched in all other labrisomids but branched in the Tripterygiidae, most Blenniidae and some Dactyloscopi-dae. None of the analytical methods used in our study support the monophyly of the Cryptotremini (Figure 4). Instead the genus Calliclinus is the sister group of all other clinioids, the northern genus Alloclinus is grouped with the southern genus Auchenionchus, and that clade is the sister group of the clinioids exclusive of the Clinidae and Calliclinus. While support values for many of these relationships are not strong, the evolutionary position of cryptotremins as early branching lineages of a diverse clade including the labrisomids was also suggested by morphological  and allozyme data . Additional study of cryptotremin blennies should provide a clearer picture of clinioid relationships.
Species allocated to three genera previously placed in the genus Labrisomus
Labrisomus Swainson, 1837
Labrisomus conditus Sazima, Carvalho-Filho, Gasparini and Sazima, 2009 a
Labrisomus cricota Sazima, Gasparini and Moura, 2002 a
Labrisomus fernandezianus (Guichenot, 1848)#
Labrisomus jenkinsi (Heller and Snodgrass, 1903)
Labrisomus multiporosus Hubbs, 1953
Labrisomus nuchipinnis (Quoy and Gaimard 1824)*+
Labrisomus philippii (Steindachner, 1866)
Labrisomus pomaspilus Springer and Rosenblatt, 1965 a
Labrisomus socorroensis Hubbs, 1953
Labrisomus wigginsi Hubbs, 1953
Labrisomus xanti Gill, 1860 +
Brockius Hubbs, 1953
Brockius albigenys (Beebe and Tee-Van, 1928)
Brockius nigricinctus (Howell Rivero, 1936) +
Brockius striatus (Hubbs, 1953)* +
Gobioclinus Gill, 1860
Gobioclinus bucciferus (Poey, 1868) +
Gobioclinus dendriticus (Reid, 1935)
Gobioclinus filamentosus (Springer, 1960)
Gobioclinus gobio (Valenciennes 1836)*
Gobioclinus guppyi (Norman, 1922) +
Gobioclinus haitiensis (Beebe and Tee-Van, 1928) +
Gobioclinus kalisherae (Jordan, 1904)
Ten of the twenty-one Malacoctenus species are included in our study and these form a well-supported clade that is the sister group of Labrisomus sensu stricto (Figure 4). Our study recovered one well-supported lineage within Malacoctenus comprising five Eastern Pacific species. However, greater taxon sampling, especially of Caribbean species, is needed to resolve the relationships within this speciose genus.
The monophyly of the tribe Starksiini, including Starksia and Xenomedea, is supported in our analysis (Figure 4). Members of this lineage reportedly share internal fertilization and a modification of the first anal-fin spine that functions as an intromittent organ in males [14, 63]. However, a recent study indicates that reproductive modes vary within this lineage and not all Starksia species exhibit internal fertilization .
The two genera currently included in the tribe Paraclinini, Paraclinus and Exerpes, share the unique characters of a spine on the posterior margin of the opercle and 0–2 segmented dorsal-fin rays [14, 65, 66]. Results from our study support the monophyly of the tribe (Figure 4), although the monotypic genus Exerpes is nested within Paraclinus, rendering the genus Paraclinus paraphyletic. Exerpes asper can be distinguished from members of the genus Paraclinus in having greatly prolonged snout, no cirri on the nape or eye, scales in the anterior segment of the lateral line consisting of a pore at each end of a tube, and by the absence of the suborbital lateral-line canal . However, this species shares several features such as an elongate snout with selected Paraclinus species, especially Paraclinus infrons. Additional study of relationships within this lineage are needed, but based on our findings and those of Brooks , we synonomize the genus Exerpes Jordan and Evermann, 1896 with Paraclinus Mocquard, 1888.
The phylogenetic relationships of the seven species of small (< 55 mm SL), cryptic, eel-like blennioids of the genus Stathmonotus have been controversial for many years . Jordan  considered Stathmonotus to be closely related to chaenopsids based on both groups lacking scales. However, Springer  included the scaled species Auchenistius stahli in Stathmonotus, and suggested that their affinities were with Paraclinus, not chaenopsids. Consequently, Stathmonotus was included in the generalized blenniiform family Labrisomidae [68, 69]. Based on six morphological synapomorphies, Hastings and Springer  placed Stathmonotus back in the Chaenopsidae as the sister group of the Chaenopsinae (= Chaenopsidae of Stephens, 1963; ). The present study based on molecular data confirms the distinctiveness of Stathmonotus, but the BI analysis supports its relationship as the sister group of the Paraclinini (Figure 4). Stathmonotus is similar to Paraclinus in post-cranial morphology, especially in having the dorsal fin comprised entirely (or mostly in the case of some Paraclinus species) of robust spines. However, it is similar to chaenopsids in cranial morphology and parsimony analysis of morphological characters places Stathmonotus with the chaenopsids because more characters are evident within the cranial region . These observations, with different suites of characters shared with disparate lineages, are consistent with a hybrid origin for this enigmatic taxon , but this hypothesis remains speculative, requiring further study. Given its morphological distinctiveness , we recommend recognizing the tribe Stathmonotini within the Labrisomidae.
Stephens  hypothesized the close relationship of the temperate hole-dwelling genus Neoclinus with the tropical chaenopsid blennies. This relationship was given further credence by the discovery of the tropical species Mccoskerichthys sandae that resembles Neoclinus, but also shares several morphological features (e.g., medially fused nasal bones) with chaenopsids . The recent placement of Neoclinus and Mccoskerichthys in the lineage with the chaenopsids based on a parsimony analysis of morphological features  is not supported in our study and implies significant morphological convergence associated with the hole-dwelling lifestyles of these fishes. The sister-group relationship between Neoclinus and Mccoskerichthys is well supported in our study, thus we recommend their designation as the Neocliniini. However, their placement as the sister group of the clinioids exclusive of the Cryptotremini and Clinidae (Figure 4) is only weakly supported by Bayesian Inference. Further study is required to determine their position within the Blenniiformes.
In our analysis, the Chaenopsidae is monophyletic with the exclusion of Stathmonotus, Neoclinus and Mccoskerichthys, three genera allocated to this group by Hastings and Springer . This result resurrects the definition of the Chaenopsidae sensu Stephens [70, 73]. This clade is well-supported by morphological features and relationships within it were recently analyzed by Lin and Hastings  using a combination of molecular and morphological characters. Results from the present analysis that included a much broader selection of outgroup taxa but fewer ingroup species, are largely congruent with those of Lin and Hastings  except for species relationships within Acanthemblemaria, a group recently studied in greater detail .
The dactyloscopids, also known as sand stargazers, are a distinctive group with several known morphological syanpomorphies, many of which are associated with their sand or gravel dwelling behaviors [1, 11, 36]. In our study, only four out of the 48 dactyloscopid species were included and they formed a well-supported monophyletic group based on our molecular data (Figure 4).
Unresolved issues in blenniiform relationships
While our analysis clarifies several aspects of the relationships of the Blenniiformes, several significant questions remain. Relationships within the Tripterygiidae are not well understood and our study, with limited taxon sampling, contributes little to this issue. Inclusion of taxa from the southern Pacific Ocean, especially from New Zealand and southern Australia where this group is especially diverse [24, 39–42], is needed to resolve relationships within the triplefins. Within the Blenniidae, the relationships of the low diversity, but morphologically distinctive, tribes Blenniini and Phenablenniini need further study. Also, the monophyly of the Parablenniini has not been confirmed although its reality and its relationships to other blenniids may have significant bearing on relationships within the combtooth blennies . Several significant questions remain regarding relationships among the lineages here termed the clinioids (node I; Figures 1 and 4). Chief among these is the relationships of the species currently allocated to the Cryptotremini. This low diversity group of relatively generalized blennies is apparently not monophyletic and in our analysis, its members are variously placed as the sister group of speciose clades of blennies. Morphological convergence in lifestyle and associated morphological features appear to have confounded past morphologically-based analyses of blenny relationships, especially among the tube-dwelling lineages. Also, relationships of the enigmatic worm blennies, genus Stathmonotus, remain unclear as they have for decades. This study supports other recent studies [32, 37] in hypothesizing a sister-group relationship of the Dactyloscopidae and Chaenopsidae (node III; Figure 4). Preliminary study indicates possible morphological features uniting these two (PAH, personal observations), but characterization of this clade may have been confounded by the inclusion of the apparently unrelated fishes of the genera Neoclinus, Mccoskerichthys and Stathmonotus within the Chaenopsidae. Finally, our understanding of species level-relationships within most lineages included herein suffers from poor taxon sampling. Thus relationships recovered in our analyses within speciose genera for which we had few representative species should be considered tentative.
Because of these and other significant issues, we are reluctant to recommend major changes to the higher-level nomenclature of the Blenniiformes at this time. While progress has been made, these and other remaining challenges in resolving blenniiform relationships will require incorporation of additional molecular markers for these and additional taxa, and importantly inclusion of morphological features in a total evidence analysis.
Biogeography of the Blenniiformes
Our mean estimated divergence time for the Blenniiformes, 60.3 mya (Additional file 2: Figure S1), is similar to recent estimates from other studies of ray-finned fishes that place the origin of the Blenniiformes at around 60  to 68 mya . Our study also implies the origin of a largely Neotropical clade of over 240 species (node IV) of the Labrisomidae, Chaenopsidae and Dactyloscopidae at approximately 37.6 mya, somewhat more recent than other estimates of 40 mya  and 48 mya  for the same clade.
The majority of species in the families Blenniidae and Tripterygiidae occur in the tropical IndoPacific with various other lineages of these families found in the Neotropics and temperate waters of the world [6, 77, 78]. Our analysis implies an origin of the Blenniiformes in the present-day IndoPacific region (Figure 5, Additional file 3: Figure S2), consequently in the Tethys Sea at approximately 60 mya. The origin of the largely Neotropical clade including the Labrisomidae, Chaenopsidae and Dactyloscopidae (node IV) at approximately 37.6 mya is coincident with the increasing opening of the Atlantic Ocean and separation of the New World and Old World land masses, but well before the closing of the Tethys Sea corridor at 12–18 mya [79, 80]. This result supports the findings of Bellwood and Wainwright  that the east and west Tethyan reef-fish faunas had diverged well before the terminal Tethyan event.
Interestingly, the intervening blenniiform clades between the primarily east and west Tethyan groups are almost exclusively temperate in distribution (Figure 5, Additional file 3: Figure S2). This finding implies either an evolutionary pathway across temperate coastal areas between increasingly separate east and west tropical regions of the Tethys Sea, or more likely, the subsequent restriction of the intervening clades to temperate refugia . These groups include low-diversity lineages currently included in the paraphyletic Cryptotremini (e.g., Calliclinus, Auchenionchus, Alloclinus and Cryptotrema)  and the genus Neoclinus[6, 61], as well as the relatively speciose Clinidae, with 85 species, that has undergone significant diversification within distant temperate regions of the world . The occurrence of these related clades in both northern and southern temperate regions indicates that they have successfully crossed intervening tropical regions during their evolutionary history [82, 83]. Better resolution of relationships and timing of divergence of these largely temperate lineages [6, 77, 78] is an important key to reconstructing the biogeographic history of the Blenniiformes in greater detail.
Other Neotropical lineages of blenniiforms include the salariin blenniid genera Ophioblennius and Scartichthys (the latter not included in our study), part or parts of the Parablenniini [48, 84], and one or more lineages within the Tripterygiidae [6, 24], as well as the Gobiesocini within the sistergroup of blenniiforms, the Gobiesocidae . A more fully resolved phylogeny of these groups with increased taxon sampling is needed to determine which, if any, of these lineages may have diverged coincident with the Neotropical blenniiforms (node IV). A few species of blennies from other primarily IndoPacific clades have invaded the Neotropics including species of the salariin genus Entomacrodus and a single species of the nemophin blenniid genus Plagiotremus. These species are well-nested within IndoPacific clades (Figure 5), supporting the hypotheses that they dispersed to the Neotropics after the origin of their respective genera in the IndoPacific region . Similarly, the few species of primarily Neotropical clades occuring in the eastern Atlantic (e.g., species of the labrisomid genera Labrisomus and Malacoctenus; [6, 87]), and members of the temperate genus Neoclinus in the northwestern Pacific , likely represent relatively recent dispersal events from their regions of origin.
In this study we reconstruct the phylogeny of the Blenniiformes with significantly broader taxon sampling (150 blenniiform species), substantially more genetic information (one mitochondrial and four nuclear markers), and more strategic outgroup representation (Gobiesocidae, Opistognathidae, and Grammatidae) than in previous studies. Progress has been made in resolving the blenniiform evolutionary relationships especially at the inter-familial level and within the Labrisomidae. Several nomenclatural changes are proposed especially in the “clinioid” clade. Examination of global distributions of blenny genera included in our analysis and estimation of divergence times imply an origin of the Blenniiformes in the present-day IndoPacific region, consequently in the Tethys Sea around 60 million years ago. A large and diverse Neotropical clade (node IV, Figures 1, 4, and 5) arose around 37.6 million years ago with intervening lineages largely restricted to temperate regions.
Molecular data for 158 terminal taxa were collected. Additional file 1: Table S1 details the included species, collection localities and deposition of voucher specimens. Our taxon sampling included representatives of all six blenniiform families (Table 1): Tripterygiidae (18 species), Blenniidae (48), Labrisomidae (36), Chaenopsidae (30), Clinidae (14), and Dactyloscopidae (4), as well as the outgroups Gobiesocidae (6), Opistognathidae (1), and Grammatidae (1). Currently recognized lineages within the blenniiform families were sampled with representative species where available (Table 1). Tissue samples were from the Marine Vertebrate Collection at Scripps Institution of Oceanography (SIO), the University of Kansas Natural History Museum (KU), the Biodiversity Research Museum at Academia Sinica (ASIZP), Taiwan and the Australian Museum (AM).
DNA extraction, amplification and sequencing
Total genomic DNA was extracted from muscle tissue with a Qiagen (Chatsworth, CA) QIAquick Tissue Kit following the manufacturer’s instructions. DNA sequences of one mitochondrial DNA marker, Cytochrome C Oxidase I (COI), and four nuclear markers, TMO-4C4, RAG1, Rhodopsin and Histone H3, were obtained. In addition to the primers used in a recent publication on chaenopsid phylogeny , six new primers were designed for amplifying PCR products across this broad sample of taxa: two extended inside primers from TMO-F3 and TMO-R3 for TMO-4C4, TMO-F4 5′-GGTGAAGTGGTTCTGCAACA-3′ and TMO-R4 5′-GCYGTGTACTCNGGRATRGT-3′; two gobiesocid-specific inside primers for RAG1, Rag-GoF 5′-TTCCTCGATCATTTAGTTTCCA-3′ and Rag-GoR 5′-GAAGGGCTTGGAGGAAACTC-3′; two blenniiform-specific inside primers for Rhodopsin, Rhod-BleF 5′-CGTCACCCTCGAACACAAGAA-3′ and Rhod-BleR 5′-GTTGTAGATGGAGGAACTCTT-3′. The PCR was performed on a Mastercycler EP Gradient S (Eppendorf, Hamburg, Germany) with the following conditions: 94°C for one minute for initial denaturing, 35 cycles of 94°C for 30 sec, 52-56°C for 45 seconds, and 72°C for 45 sec, followed by 72°C for 5 minutes as the final extension. Resulting amplicons were purified with Exonuclease I (20 U/μl, New England Biolabs) and Shrimp Alkaline Phosphatase (1 U/μl, Roche) in order to remove single-stranded DNA and unincorporated dNTPs. Sequencing was done in both directions with the amplification primers and DYEnamicTM ET dye terminator sequencing kit on an automated MegaBACE™ 500 DNA sequencer (Amersham Biosciences Corp., Piscataway, NJ).
DNA sequence alignment, partition, and analysis
Sequences were assembled, edited, and aligned with no gaps interrupting reading frames based on translated protein sequences in Geneious . Nucleotide sequences were checked on the NCBI database (http://www.ncbi.nlm.nih.gov/) for possible stop codons as an indication of pseudogenes. Because all five genetic markers are protein-coding genes, the rapidly evolving third codon positions were partitioned from the slower evolving first and second positions in Bayesian and likelihood analyses, resulting in ten total partitions . To avoid base frequency deviations across taxa which can potentially mislead phylogenetic reconstruction , the chi-square test of base frequency homogeneity across taxa was executed for each partition using PAUP 4.0b10 . The Akaike Information Criterion (AIC)  implemented in jModelTest v0.1.1 [94, 95] was used to select the best-fit evolutionary model for each partition (Table 2). Likelihood calculations were carried out for 88 models, including 11 substitution schemes, equal or unequal base frequencies, a proportion of invariant sites (I), and rate variation among sites with four rate categories (G) on a BIONJ-JC fixed tree.
Maximum likelihood (ML), Bayesian inference (BI), and maximum parsimony (MP) analyses were conducted to reconstruct the phylogenetic relationships. Maximum likelihood tree searching was conducted in Garli 2.0  under the CIPRES Science Portal v2.2 . Best-fit evolutionary models selected by jModeltest were applied to the ten genetic partitions (five genes for first + second and third codon positions) (Table 2). If a selected model could not be implemented in Garli, the least complex model that included all of the parameters of the selected model was used instead. Therefore, the General Time Reversible model with gamma rate heterogeneity model (GTR+G) was selected for the third codon position of COI, and the GTR+I+G model  was selected for all other partitions. Twelve replicates were run to find the tree topology with the best likelihood. The setting of 1,000 replicate bootstrap analysis was identical to the above, except the number of generations without topology improvement required for termination (genthreshfortopoterm) was reduced from 20,000 (default) to 10,000 to reduce the running time as suggested in the manual .
Bayesian Metropolis coupled Markov chain Monte Carlo (MCMC) estimation of the phylogeny was carried out using MrBayes v3.1.2  under the CIPRES Science Portal. The dataset was partitioned and assigned evolutionary models as in the ML analyses. A partitioned mixed-model analysis was applied and all model parameter values were “unlinked” among partitions . In all analyses, the average substitution rates (prset ratepr = variable) and model parameters including the branch lengths within the tree (unlink brlens) were allowed to vary among partitions. Two simulated independent runs were performed for 10 million generations each and starting from different random trees. Each run comprised four chains (one cold and three heated) and was sampled every 1,000 generations. The sampled parameter values from Bayesian MCMC were evaluated in Tracer v1.4 (http://beast.bio.ed.ac.uk/Tracer) and the generations before reaching a plateau were discarded as burnin. Trees from the stationary phase of the two runs were then pooled by LogCombiner v1.5.4  and the 50% majority tree was exported by Mesquite v2.73 . Assigning this tree as the target tree, the posterior probability of each node and the mean branch lengths were calculated with TreeAnnotator v1.5.4 .
Maximum parsimony analyses were conducted using PAUP 4.0b10. Heuristic searches were performed using tree bisection reconnection (TBR) with branch-swapping from 1,000 random-addition-sequence replicates to avoid entrapment in local optima. All nucleotide sites were equally weighted and gaps were treated as missing characters. Nonparametric node supports for trees were estimated with 1,000 heuristic searches (maxTree = 500) starting with 10 random addition sequence replicates.
Divergence times of the blenniiform lineages sampled in our study were estimated with a relaxed molecular clock analysis . Relative divergence times of nodes were estimated in BEAST v1.7.2 assuming an uncorrelated lognormal model (UCLN) of rate variation among branches in the tree and a Yule-process-speciation prior of the branching rate [101, 103]. As in the maximum likelihood and Bayesian analysis described before, the same ten data partitions and molecular evolutionary models were applied to estimate the posterior density of relative divergence times. With TreeEdit v1.0a10 (http://tree.bio.ed.ac.uk/software/treeedit/), the best ML tree from Garli was converted to an ultrametric tree assuming the origin of the Blenniidae as 41 million years old as estimated from a bony fish phylogeny including 21 molecular markers and 1410 fish taxa  and assigned as the starting tree. We did not incorporate the putative blenniid fossil from the Monte Bolca formation (50 mya; ), because its identity as a blenniid is in doubt ([11, 105], Springer, pers. comm). Three independent MCMC analyses were each run for 40 million generations, sampled every 1000 generations, and discarding the first 10% of samples, resulting in acceptable mixing as determined by Tracer. These three runs were combined to obtain an estimate of the posterior distribution. The posterior probability density of the combined tree and log files was summarized on a maximum clade credibility tree with TreeAnnotator.
The general distributions of blenniiform genera with representatives included in our analysis were scored as either tropical IndoPacific (I), Neotropical (N), or temperate (T; Additional file 4: Table S2). Genera with species present in more than one of these were scored as polymorphic except where phylogenetic evidence implies recent dispersal into another region after the origin of the genus. For example, Plagiotremus (Blenniidae) was scored as IndoPacific because the single Neotropical species (P. azaleus) is nested within an otherwise entirely IndoPacific clade ([6, 47]; Figure 3).
We used both parsimony implemented in Mesquite and Bayesian Binary MCMC (BBM) methods implemented in Reconstruct Ancestral States in Phylogenetics v2.1 beta (RASP; ) to reconstruct geographical areas at nodes in the phylogeny. The genus–level topology of the best ML tree was used for the reconstruction. The BBM analysis was conducted with estimated character state frequencies (F81) and gamma-distributed rate variation between sites. The root distribution was set to null (i.e. the outgroup is assigned to a new area where none of the ingroup taxa occur). Two independent runs of 10 chains with a temperature of 0.1 were run for one million generations, sampled every 100 generations, and the first 2,500 samples were discarded. A distance between runs of less than 0.01 was used as an indication of convergence.
We thank N. Knowlton for providing laboratory space and equipment, G. Rouse for assisting with data analyses and providing critical comments, R. Burton for insightful comments on molecular methods and the sequencing machine, HJ Walker for curatorial assistance, and R. Rosenblatt and C. Wills for providing valuable comments. Tissue samples were generously provided by the Marine Vertebrate Collection at Scripps Institution of Oceanography, University of Kansas Natural History Museum, Biodiversity Research Museum at Academia Sinica, Taiwan and the Australian Museum. This work was supported by the Study-abroad Scholarship from the Ministry of Education, Taiwan, SIO Rosenblatt Fellowship, the E. B. Scripps Foundation, a gift from B. Shor, the UCSD Academic Senate and a dissertation improvement grant from UC MEXUS.
- Hastings PA, Springer VG: Systematics of the Blennioidei and the included families Dactyloscopidae, Chaenopsidae, Clinidae and Labrisomidae. The Biology of Blennies. Edited by: Patzner RA, Gonçalves EJ, Hastings PA, Kapoor BG. 2009, Enfield: Science Publishers, 3-30.Google Scholar
- Griffiths SP: Rockpool ichthyofaunas of temperate Australia: species composition, residency and biogeographic patterns. Estuar Coast Shelf Sci. 2003, 58: 173-186. 10.1016/S0272-7714(03)00073-8.Google Scholar
- Thomson DA, Gilligan MR: Rocky-shore Fishes. A new island biogeography of the Sea of Cortes. Edited by: Case TJ, Cody ML, Ezcurra E. 2002, New York: Oxford University Press, 154-180.Google Scholar
- Paulin C, Roberts C: The Rockpool Fishes of New Zealand. 1992, Wellington: Museum of New Zealand Te Papa TongarewaGoogle Scholar
- Patzner RA, Gonçalves EJ, Hastings PA, Kapoor BG: The Biology of Blennies. 2009, Enfield: Science PublishersGoogle Scholar
- Hastings PA: Biogeography of New World Blennies. The Biology of Blennies. Edited by: Patzner RA, Gonçalves EJ, Hastings PA, Kapoor BG. 2009, Enfield: Science Publishers, 95-118.Google Scholar
- Rosenblatt RH: Some aspects of speciation in marine shore fishes. Syst Assoc Publ. 1963, 5: 171-180.Google Scholar
- Hastings PA: Biogeography of the tropical eastern Pacific: distribution and phylogeny of chaenopsid fishes. Zool J Linn Soc. 2000, 128: 319-335. 10.1111/j.1096-3642.2000.tb00166.x.Google Scholar
- Stepien CA, Dillon AK, Brooks MJ, Chase KL, Hubers AN: The evolution of blennioid fishes based on an analysis of mitochondrial 12S rRNA. Molecular Systematics of Fishes. Edited by: Kocher TD, Stepien CA. 1997, San Diego: Academic Press, 245-270.Google Scholar
- Rosenblatt RH: Ontogeny and Systematics of Fishes. 1984, American Society of Ichthyologists and Herpetologists, 551-552. Blennioidei: Introduction, Special publication 1,Google Scholar
- Springer VG: Definition of the suborder Blennioidei and its included families (Pisces: Perciformes). Bull Mar Sci. 1993, 52: 472-495.Google Scholar
- Springer VG: Osteology and classification of fishes of the family Blenniidae. U S Natl Mus Bull. 1968, 284: 1-85.Google Scholar
- Gosline WA: The suborders of perciform fishes. Proc U S Natl Mus. 1968, 124: 1-78.Google Scholar
- Hubbs C: A contribution to the classification of the blennioid fishes of the family Clinidae, with a partial revision of eastern Pacific forms. Stanford Ichthy Bull. 1952, 4: 41-165.Google Scholar
- Jordan DS: A classification of fishes, including families and genera as far as known. Stanford Univ Publ Biol Sci. 1923, 3: 77-243.Google Scholar
- Regan CT: The classification of the blennioid fishes. Ann Mag Nat Hist Ser 8. 1912, 36: 265-280.Google Scholar
- Wiley E, Johnson G: A teleost classification based on monophyletic groups. Origin and Phylogenetic Interrelationships of Teleosts. Edited by: Nelson J, Schultze H-P, Wilson M. 2010, München: Verlag Dr. Friedrich Pfeil, 123-182.Google Scholar
- Johnson GD: Percomorph phylogeny: A survey of acanthomorphs and a new proposal. Bull Mar Sci. 1993, 52: 554-626.Google Scholar
- Springer VG, Orrell TM: Phylogenetic analysis of the families of acanthomorph fishes based on dorsal gill-arch muscles and skeleton. Bull Biol Soc Washington. 2004, 11: 237-260.Google Scholar
- Mooi RD, Gill AC: Association of epaxial musculature with dorsal-fin pterygiophores in acanthomorph fishes, and its phylogenetic significance. Bull Nat Hist Mus London, Zool. 1995, 62: 121-137.Google Scholar
- Böhlke JE, Robins CH: Description of a new genus and species of clinid fish from the Western Caribbean, with comments on the families of the Blennioidea. Proc Acad Nat Sci Philad. 1974, 126: 1-8.Google Scholar
- Stepien CA, Dixon MT, Hills DM: Evolutionary relationships of the blennioid fish families Clinidae, Labrisomidae and Chaenopsidae: congruence between DNA seqeunce and allozyme data. Bull Mar Sci. 1993, 52: 496-515.Google Scholar
- Hastings PA, Springer VG: Systematics of the Blenniidae (Combtooth Blennies). The Biology of Blennies. Edited by: Patzner RA, Gonçalves EJ, Hastings PA, Kapoor BG. 2009, Enfield: Science Publishers, 69-91.Google Scholar
- Fricke R: Systematics of the Tripterygiidae (triplefins). The Biology of Blennies. Edited by: Patzner RA, Gonçalves EJ, Hastings PA, Kapoor BG. 2009, Enfield: Science Publishers, 31-67.Google Scholar
- Chen W-J, Bonillo C, Lecointre G: Repeatability of clades as a criterion of reliability: a case study for molecular phylogeny of Acanthomorph (Teleostei) with larger number of taxa. Mol Phyl Evol. 2003, 26: 262-288. 10.1016/S1055-7903(02)00371-8.Google Scholar
- Simmons MP, Miya M: Efficiently resolving the basal clades of a phylogenetic tree using Bayesian and parsimony approaches: a case study using mitogenomic data from 100 higher teleost fishes. Mol Phyl Evol. 2004, 31: 351-362. 10.1016/j.ympev.2003.08.004.Google Scholar
- Miya M, Satoa TP, Nishida M: The phylogenetic position of toadfishes (order Batrachoidiformes) in the higher ray-finned fish as inferred from partitioned Bayesian analysis of 102 whole mitochondrial genome sequences. Biol J Linn Soc. 2005, 85: 289-306. 10.1111/j.1095-8312.2005.00483.x.Google Scholar
- Dettai A, Lecointre G: Further support for the clades obtained by multiple molecular phylogenies in the acanthomorph bush. C R Biol. 2005, 328: 674-689. 10.1016/j.crvi.2005.04.002.PubMedGoogle Scholar
- Setiamarga DHE, Miya M, Yamanoue Y, Mabuchi K, Satoh TP, Inoue JG, Nishida M: Interrelationships of Atherinomorpha (medakas, flyingfishes, killifishes, silversides, and their relatives): The first evidence based on whole mitogenome sequences. Mol Phyl Evol. 2008, 49: 598-605. 10.1016/j.ympev.2008.08.008.Google Scholar
- Li B, Dettaï A, Cruaud C, Couloux A, Desoutter-Meniger M, Lecointre G: RNF213, a new nuclear marker for acanthomorph phylogeny. Mol Phyl Evol. 2009, 50: 345-363. 10.1016/j.ympev.2008.11.013.Google Scholar
- Smith WL, Craig MT: Casting the percomorph net widely: The importance of broad taxonomic sampling in the search for the placement of serranid and percid fishes. Copeia. 2007, 2007: 35-55. 10.1643/0045-8511(2007)7[35:CTPNWT]2.0.CO;2.Google Scholar
- Wainwright PC, Smith WL, Price SA, Tang KL, Sparks JS, Ferry LA, Kuhn KL, Eytan RI, Near TJ: The evolution of pharyngognathy: A phylogenetic and functional appraisal of the pharyngeal jaw key innovation in labroid fishes and beyond. Syst Biol. 2012, 61: 1001-1027. 10.1093/sysbio/sys060.PubMedGoogle Scholar
- Smith WL, Wheeler WC: Polyphyly of the mail-cheeked fishes (Teleostei: Scorpaeniformes): Evidence from mitochondrial and nuclear sequence data. Mol Phyl Evol. 2004, 32: 627-647. 10.1016/j.ympev.2004.02.006.Google Scholar
- Betancur-R R, Broughton R, Wiley E, Carpenter K, López J, Li C, Holcroft N, Arcila D, Sanciangco M, Cureton JI, et al: The tree of life and a new classification of bony fishes. PLOS Curr Tree of Life. 2013, Edition 1. Doi: 10.1371/currents.tol.53ba26640df0ccaee75bb165c8c26288Google Scholar
- Near TJ, Dornburg A, Eytan RI, Keck BP, Smith WL, Kuhn KL, Moore JA, Price SA, Burbrink FT, Friedman M, et al: Phylogeny and tempo of diversification in the superradiation of spiny-rayed fishes. Proc Natl Acad Sci. 2013, 110: 12738-12743. 10.1073/pnas.1304661110.PubMed CentralPubMedGoogle Scholar
- Doyle KD: Phylogeny of the sand stargazers (Dactyloscopidae: Blennioidei). Copeia. 1998, 1998: 76-96. 10.2307/1447703.Google Scholar
- Hastings PA, Springer VG: Review of Stathmonotus with redefinition and phylogenetic analysis of the Chaenopsidae (Teleostei:Blennioidei). Smith Contrib Zool. 1994, 558: 1-48.Google Scholar
- Walsh HE, Kidd MG, Moum T, Friesen VL: Polytomies and the power of phylogenetic inference. Evolution. 1999, 53: 932-937. 10.2307/2640732.Google Scholar
- Fricke R: Tripterygiid Fishes of Australia, New Zealand and the Southwest Pacific Ocean (Teleostei). 1994, Koenigstein: Koeltz Scientific BooksGoogle Scholar
- Carreras-Carbonell J, Macpherson E, Pascual M: Rapid radiation and cryptic speciation in Mediterranean triplefin blennies (Pisces: Tripterygiidae) combining multiple genes. Mol Phyl Evol. 2005, 37: 751-761. 10.1016/j.ympev.2005.04.021.Google Scholar
- Geertjes GJ, Kamping A, van Delden W, Videler JJ: Genetic relationships among one non-endemic and two endemic Mediterranean triplefin blennies (Pisces, Blennioidei). Mar Ecol. 2001, 22: 255-265. 10.1046/j.1439-0485.2001.01738.x.Google Scholar
- Hickey AJR, Clements KD: Genome size evolution in New Zealand triplefin fishes. J Hered. 2005, 96: 356-362. 10.1093/jhered/esi061.PubMedGoogle Scholar
- Loytynoja A, Goldman N: Phylogeny-aware gap placement prevents errors in sequence alignment and evolutionary analysis. Science. 2008, 320: 1632-1635. 10.1126/science.1158395.PubMedGoogle Scholar
- Williams JT: Phylogenetic relationships and revision of the blenniid fish Scartichthys. Smith Contrib Zool. 1990, 492: 1-30.Google Scholar
- Norman JR: Notes on the blennioid fishes. I. A provisional synopsis of the genera of the family Blenniidae. Ann Mag Nat Hist Ser 11. 1943, 10: 793-812.Google Scholar
- Bath H: Revision der Blenniini. Senck Biol. 1977, 57: 167-234.Google Scholar
- Smith-Vaniz WF: The saber-toothed blennies, tribe Nemophini (Pisces; Blenniidae). Monogr Acad Nat Sci Philad. 1976, 19: 1-196.Google Scholar
- Bock M, Zander CD: Osteological characters as tool for blenniid taxonomy - A generic revision of European Blenniidae (Percomorphi; Pisces). Z Zool Syst Evol. 1986, 24: 138-143.Google Scholar
- Bath H: Osteology and morphology of fishes of the subfamily Salariinae and its junior synonym Parablenniinae (Pisces: Blennidae). Stutt Beitr Naturkd. 2001, A268: 1-42.Google Scholar
- Smith-Vaniz WF, Springer VG: Synopsis of the tribe Salariini, with description of five new genera and three new species (Pisces: Blenniidae). Smith Contrib Zool. 1971, 73: 1-72.Google Scholar
- Springer VG, Williams JT: The Indo-Pacific blenniid fish genus Istiblennius reappraised: a revision of Istiblennius, Blenniella, and Paralticus, new genus. Smith Contrib Zool. 1994, 565: 1-193.Google Scholar
- Bath H: Verwandtschaftliche beziehungen der arten der gattung Hypsoblennius Gill 1861. Senck Biol. 2000, 80: 175-187.Google Scholar
- George A, Springer VG: Revision of the clinid fish tribe Ophiclinini, including five new species, and definition of the family Clinidae. Smith Contrib Zool. 1980, 307: 1-31.Google Scholar
- Stepien CA, Rosenblatt RH: Patterns of gene flow and genetic divergence in the Northeastern Pacific Clinidae (Teleostei: Blennioidei), based on allozyme and morphological data. Copeia. 1991, 1991: 873-896. 10.2307/1446084.Google Scholar
- Hubbs C: Additional records of clinid fishes, with the description of a new species of Cryptotrema from the Gulf of California. Copeia. 1954, 1: 17-19.Google Scholar
- McCosker JE, Stephens JS, Rosenblatt RH: Cottoclinus canops, a new genus and species of blenny (Perciform: Labrisomidae) from the Galápagos Islands. Proc Cal Acad Sci. 2003, 54: 155-160.Google Scholar
- Böhlke JE, Springer VG: A new genus and species of fish (Nemaclinus atelestos) from the western Atlantic (Perciformes: Clinidae). Proc Acad Nat Sci Philad. 1975, 127: 57-61.Google Scholar
- Baldwin C, Robertson R: A new Haptoclinus blenny (Teleostei, Labrisomidae) from deep reefs off Curaçao, southern Caribbean, with comments on relationships of the genus. ZooKeys. 2013, 306: 71-81.PubMedGoogle Scholar
- Stephens JS, Springer VG: Clinid fishes of Chile and Peru, with description of a new species, Myxodes ornatus, from Chile. Smith Contrib Zool. 1974, 159: 1-23.Google Scholar
- Springer VG: Systematics and zoogeography of the clinid fishes of the subtribe Labrisomini Hubbs. Publ Inst Mar Sci Univ Tex. 1959, 5: 417-492.Google Scholar
- Hubbs C: Revision and systematic position of blenniid fishes of the genus Neoclinus. Copeia. 1953, 1953: 11-23. 10.2307/1440237.Google Scholar
- Cuvier G, Valenciennes A: Histoire naturelle des poissons. 1836, Paris: De l’imprimerie de Regnaud ChaudiereGoogle Scholar
- Rosenblatt RH, Taylor LR: The Pacific species of the clinid fish tribe Starksiini. Pac Sci. 1971, 25: 436-463.Google Scholar
- Fishelson L, Baldwin CC, Hastings PA: Gonad morphology, gametogenesis, and reproductive modes in fishes of the tribe Starksiini (Teleostei, Blenniiformes). J Morph. 2013, 274: 496-511. 10.1002/jmor.20110.PubMedGoogle Scholar
- Rosenblatt RH, Parr TD: The Pacific species of the clinid fish genus Paraclinus. Copeia. 1969, 1969: 1-20. 10.2307/1441691.Google Scholar
- Brooks MJ: The ontogeny and evolution of sexual dimorphism in paraclinin blennies (Teleostei, Labrisomidae). 1992, Tucson: University of ArizonaGoogle Scholar
- Springer VG: The taxonomic status of fishes of the genus Stathmonotus, including a review of the Atlantic species. Bull Mar Sci. 1955, 5: 66-80.Google Scholar
- Eschmeyer WN: Catalog of the Genera of Recent fishes. 1990, San Francisco: California Academy of SciencesGoogle Scholar
- Nelson JS: Fishes of the World. 1984, New York: John Wiley & Sons, 2Google Scholar
- Stephens JS: A revised classification of the blennioid fishes of the American family Chaenopsidae. Univ California Publ Zool. 1963, 68: 1-165.Google Scholar
- McDade LA: Hybridization and phylogenetics: Special insights from morphology. Phylogenetic Analysis of Morphological Data. Edited by: Wiens J. 2000, Washington DC: Smithsonian Institution Press, 146-164.Google Scholar
- Rosenblatt R, Stephens J: Mccoskerichthys sandae, a new and unusual chaenopsid blenny from the Pacific coast of Panama and Costa Rica. Los Angeles County Mus Contrib Sci. 1978, 293: 1-22.Google Scholar
- Stephens JS: Seven new chaenopsid blennies from the western Atlantic. Copeia. 1970, 1970: 280-309. 10.2307/1441651.Google Scholar
- Lin H-C, Hastings PA: Evolution of a Neotropical marine fish lineage (Subfamily Chaenopsinae, Suborder Blennioidei) based on phylogenetic analysis of combined molecular and morphological data. Mol Phyl Evol. 2011, 60: 236-248. 10.1016/j.ympev.2011.04.018.Google Scholar
- Eytan RI, Hastings PA, Holland BR, Hellberg ME: Reconciling molecules and morphology: Molecular systematics and biogeography of Neotropical blennies (Acanthemblemaria). Mol Phyl Evol. 2012, 62: 159-173. 10.1016/j.ympev.2011.09.028.Google Scholar
- Near TJ, Eytan RI, Dornburg A, Kuhn KL, Moore JA, Davis MP, Wainwright PC, Friedman M, Smith WL: Resolution of ray-finned fish phylogeny and timing of diversification. Proc Natl Acad Sci. 2012, 109: 13698-13703. 10.1073/pnas.1206625109.PubMed CentralPubMedGoogle Scholar
- Almada VC, Robalo JI, Gonçalves EJ, Levy A, Patzner RA: Blennies in temperate seas. The Biology of Blennies. Edited by: Patzner RA, Gonçalves EJ, Hastings PA, Kapoor BG. 2009, Enfield: Science Publishers, 119-135.Google Scholar
- Stepien CA: Evolution and biogeography of the Clinidae (Teleostei: Blennioidei). Copeia. 1992, 1992: 375-392. 10.2307/1446198.Google Scholar
- Dercourt J, Zonenshain LP, Ricou LE, Kazmin VG, Le Pichon X, Knipper AL, Grandjacquet C, Sbortshikov IM, Geyssant J, Lepvrier C, et al: Geological evolution of the Tethys belt from the Atlantic to the Pamirs since the Lias. Tectonophysics. 1986, 123: 241-315. 10.1016/0040-1951(86)90199-X.Google Scholar
- Adams C, Gentry A, Whybrow P: Dating the terminal Tethyan event. Reconstruction of Marine Paleoenvironments. Edited by: Meulenkamp JE. 1983, Hoogeveen: Utrecht Micropalaeontological Bulletins, 273-298.Google Scholar
- Bellwood DR, Wainwright PC: The history and biogeography of fishes on coral reefs. Coral Reef Fishes. Edited by: Peter FS. 2002, San Diego: Academic Press, 5-32.Google Scholar
- Briggs JC, Bowen BW: Marine shelf habitat: Biogeography and evolution. J Biogeogr. 2013, 40: 1023-1035. 10.1111/jbi.12082.Google Scholar
- Burridge C: Antitropicality of Pacific fishes: Molecular insights. Env Biol Fish. 2000, 65: 151-164.Google Scholar
- Levy A, von der Heyden S, Floeter SR, Bernardi G, Almada VC: Phylogeny of Parablennius Miranda Ribeiro, 1915 reveals a paraphyletic genus and recent Indo-Pacific diversification from an Atlantic ancestor. Mol Phyl Evol. 2013, 67: 1-8. 10.1016/j.ympev.2012.12.010.Google Scholar
- Briggs JC: A monograph of the clingfishes (Order Xenopterygii). Stanford Ichthy Bull. 1955, 6: 1-224.Google Scholar
- Springer VG: Revision of the circumtropical shorefish genus Entomacrodus (Blenniidae: Salariinae). Proc U S Natl Mus. 1967, 122: 1-150.Google Scholar
- Floeter SR, Rocha LA, Robertson DR, Joyeux JC, Smith-Vaniz WF, Wirtz P, Edwards AJ, Barreiros JP, Ferreira CEL, Gasparini JL, et al: Atlantic reef fish biogeography and evolution. J Biogeogr. 2008, 35: 22-47.Google Scholar
- Fukao R: Fishes of Neoclinus bryope species complex from Shirahama, Japan, with description of two new species. Jap J Ichthy. 1987, 34: 291-308.Google Scholar
- Drummond A, Ashton B, Buxton S, Cheung M, Cooper A, Duran C, Field M, Heled J, Kearse M, Markowitz S, et al: Geneious. 5.4 edn. 2011Google Scholar
- Brandley M, Schmitz A, Reeder T: Partitioned Bayesian analyses, partition choice, and the phylogenetic relationships of scincid lizards. Syst Biol. 2005, 54: 373-390. 10.1080/10635150590946808.PubMedGoogle Scholar
- Saccone C, Pesole G, Preparata G: DNA microenvironments and the molecular clock. J Mol Evol. 1989, 29: 407-411. 10.1007/BF02602910.PubMedGoogle Scholar
- Swofford DL: PAUP*: Phylogenetic analysis using parsimony (and other methods). 4.0 Beta edn. 2002, Sunderland, MA: Sinauer AssociatesGoogle Scholar
- Akaike H: A new look at the statistical model identification. IEEE Trans Aut Contr. 1974, 19: 716-723. 10.1109/TAC.1974.1100705.Google Scholar
- Guindon S, Gascuel O: A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol. 2003, 52: 696-704. 10.1080/10635150390235520.PubMedGoogle Scholar
- Posada D: jModelTest: Phylogenetic model averaging. Mol Biol Evol. 2008, 25: 1253-1256. 10.1093/molbev/msn083.PubMedGoogle Scholar
- Zwickl DJ: Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. PhD thesis. 2006, Austin: The University of TexasGoogle Scholar
- Miller MA, Pfeiffer W, Schwartz T: Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Proceedings of the Gateway Computing Environments Workshop (GCE). New Orleans. 2010, 1-8.Google Scholar
- Tavaré S: Some probabilistic and statistical problems in the analysis of DNA sequences. Am Math Soc: Lectures on mathematics in the Life Sciences. 1986, 17: 57-86.Google Scholar
- Huelsenbeck JP, Ronquist F: MRBAYES: Bayesian inference of phylogeny. Bioinformatics. 2001, 17: 754-755. 10.1093/bioinformatics/17.8.754.PubMedGoogle Scholar
- Ronquist F, Huelsenbeck JP: MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics. 2003, 19: 1572-1574. 10.1093/bioinformatics/btg180.PubMedGoogle Scholar
- Drummond A, Rambaut A: BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol. 2007, 7: 214-10.1186/1471-2148-7-214.PubMed CentralPubMedGoogle Scholar
- Maddison WP, Maddison DR: Mesquite: a modular system for evolutionary analysis. 2.73 edn. 2011Google Scholar
- Drummond AJ, Ho SYW, Phillips MJ, Rambaut A: Relaxed phylogenetics and dating with confidence. PLoS Biol. 2006, 4: e88-10.1371/journal.pbio.0040088.PubMed CentralPubMedGoogle Scholar
- Bellwood DR: The Eocene fishes of Monte Bolca: the earliest coral reef fish assemblage. Coral Reefs. 1996, 15: 11-19.Google Scholar
- Blot J: Actinopterygii, Ordre des Scorpaeniformes? Famille des Pterygocephalidae Blot 1980. Studi Ricerche Giacimenti Terz. Bolca. 1984, 4: 265-305.Google Scholar
- Yu Y, AJ H, X-J H: RASP (Reconstruct Ancestral State in Phylogenies). 2.1 beta edn. 2013Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.