We present a multi-locus phylogenetic analysis of the shallow water (high intertidal) barnacle genus Chthamalus, focusing on member species in the western hemisphere. Understanding the phylogeny of this group improves interpretation of classical ecological work on competition, distributional changes associated with climate change, and the morphological evolution of complex cirripede phenotypes.
Methodology and Findings
We use traditional and Bayesian phylogenetic and ‘deep coalescent’ approaches to identify a phylogeny that supports the monophyly of the mostly American ‘fissus group’ of Chthamalus, but that also supports a need for taxonomic revision of Chthamalus and Microeuraphia. Two deep phylogeographic breaks were also found within the range of two tropical American taxa (C. angustitergum and C. southwardorum) as well.
Our data, which include two novel gene regions for phylogenetic analysis of cirripedes, suggest that much more evaluation of the morphological evolutionary history and taxonomy of Chthamalid barnacles is necessary. These data and associated analyses also indicate that the radiation of species in the late Pliocene and Pleistocene was very rapid, and may provide new insights toward speciation via transient allopatry or ecological barriers.
Citation: Wares JP, Pankey MS, Pitombo F, Daglio LG, Achituv Y (2009) A “Shallow Phylogeny” of Shallow Barnacles (Chthamalus). PLoS ONE4(5): e5567. https://doi.org/10.1371/journal.pone.0005567
Editor: John F. Bruno, University of North Carolina at Chapel Hill, United States of America
Received: November 30, 2008; Accepted: April 1, 2009; Published: May 15, 2009
Copyright: © 2009 Wares et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: We acknowledge financial support from the U.S.-Israel Binational Science Foundation (grant #2004/239) and the University of Georgia Research Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The high intertidal barnacles of the genus Chthamalus have a high profile of systematic and biogeographic exploration, in part due to the confusing degree of morphological plasticity and similarity among species – a trait that promoted Darwin's  interest in the taxonomy and evolution of barnacles. Additionally, the genus Chthamalus has proven to be an excellent group of organisms for experimental ecology, in particular pertaining to interspecific competition , , . As this genus tends to be found in the highest reaches of the intertidal zone, it experiences some of the most dramatic daily and seasonal physiological transitions, and is one of several taxa that can be used as a clear indicator that global climate change is forcing the redistribution of many marine species , .
The genus is globally distributed, locally abundant, and relatively diverse, with over 20 nominal species , . The distributions of these species, and heritable diversity within each, are largely determined through abiotic mechanisms because the sole mechanism of dispersal is the planktotrophic larval stage, which may require 4–6 weeks of development before an individual is competent to settle. Thus, changes in the nearshore environment and changes in nearshore ocean circulation are likely to have significant effects on the demographic isolation of populations and the diversity each population harbors. Hybridization between species is thought to be rare ; however, the morphological and ecological similarity among many of these species, with broad phenotypic variation within species and a large capacity for larval dispersal, has made distinguishing species for biogeographic or evolutionary analysis difficult. Understanding the impact of cryptic species on biogeographic patterns is critical for evaluation of what generates and maintains diversity .
Among the characters that are often used in diagnosing species in this family and genus, many are of questionable value . Pilsbry  initially separated the genus into two groups based on mandibular shape; Nilsson-Cantell  subsequently named these groups the quadridentoid Stellatus group and the tridentoid Hembeli group. Zullo  determined that the tridentoid form was the plesiomorphic state – and the closely related tropical genus Euraphia (now largely Microeuraphia, under the Euraphiinae) bears this trait – while the quadridentate genera (e.g., Chthamalus under Chthamalinae) were considered to bear the more derived trait. However, Pope argued that “the dividing line between those species with the so-called tridentate or hembeli jaw and those with the quadridentate or stellatus kind of mandible is sometimes a rather hazy one.” It is also clear that characters found on feeding appendages (cirri) that are often used to discriminate subgroups of the genus Chthamalus (e.g., Dando and Southward 1980 ) are quite variable within species and may themselves be problematic characters for discriminating the effects of phylogenetic history and current ecology , , .
With such variation in these key traits, they are nevertheless often used for taxonomic purposes . Species-level taxonomy has gone through a range of reductions  and expansions ,  as more data become available. Molecular data in particular have been invaluable for characterizing these taxa, in terms of identifying new species or lineages (e.g., , ) and phylogenetic exploration of the utility of known morphological traits for identification , .
Wares  attempted to clarify the relationships of the Tropical Eastern Pacific (TEP) species complex; although the ‘fissus group’ sensu Dando and Southward  appeared to be monophyletic, there were inadequate data to address larger phylogenetic questions, and some confusion remained about the phylogeny of the genus overall . Although molecular approaches have been quite useful for delimiting cryptic species in this group, it is difficult to find sufficient multilocus variation to identify the pattern of divergence at deeper nodes; most nuclear loci are not sufficiently variable to evaluate Miocene divergences . Because most phylogenetic studies have been regionally restricted , , we are still deficient in complete understanding of the phylogeny of this important intertidal taxon. Knowing the pattern and timing of radiation in this group may help illuminate the mechanisms involved in species divergence.
Here, we examine a broader collection of species within Chthamalus, including species from all four subgeneric groupings defined by Dando and Southward  based on the internal characters discussed above. We include closely related species in the genus Microeuraphia from the TEP and Caribbean to generate at least two pairs of geminate (across the Panamanian Isthmus) taxa for calibration of the timing of species radiation. Data from these species are used to address whether the morphological characters noted above are useful and consistent predictors of phylogeny, and what global or regional events may be temporally associated with the radiation of species along the North and Central American coasts. Most importantly, we address these questions using a variety of phylogenetic approaches, including consideration of the ‘deep coalescent’  methods that account for intraspecific variation and shared allelic variation among species to describe more precisely the radiation of species and to define portions of the phylogeny that are likely to be problematic because of rapid ancestral radiations. These methods, applied to a species group thought to have radiated in the recent Miocene , provide greater flexibility in the design of future studies to address the radiation in more detail.
Verification and Preliminary Phylogenetic Analysis
At the mitochondrial 16S locus, we obtained 499 bp of aligned sequence data for 252 individual barnacle specimens (Genbank FJ862065-306). Of these characters, 302 were constant and 24 were uninformative; 173 characters are parsimony informative. Saturation tests indicated strong phylogenetic signal at this locus, as did a plot of transitions and transversions against HKY-corrected genetic distances. Both types of substitutions showed effectively linear increasing functions with genetic distance. MP bootstrap analysis of the mt16S partition indicated a number of interior nodes with <60% bootstrap support (similar result using NJ), with a mean bootstrap value across all internal nodes of 77.4% (73.2% for NJ). All individuals identified to species with mtCOI were consistently delineated the same way with mt16S with greater than 95% bootstrap support for each taxon, with the exceptions of the two clades of C. southwardorum that were identified with mtCOI, and C. hedgecocki and C. panamensis (but which combined have bootstrap support >95% under each analysis), suggesting less phylogenetic support for those taxonomic and/or phylogeographic patterns. No individuals or data partitions were excluded from subsequent analysis based on this gene alone. The best-fit model of molecular evolution at this locus was TVM with invariant sites (p = 0.44) and gamma-distributed site rates (α = 0.76). Base frequencies were heavily A/T biased (70%); the chosen model was used for subsequent corrected distance and likelihood analysis, with a 6-parameter invariant/gamma-distributed rate set for MrBayes.
At the nuclear LTRS locus, we obtained 434 bp of aligned sequence data for 163 specimens (Genbank FJ862503-665). Of these characters, 224 were constant and 7 were uninformative; 203 characters were parsimony informative. Saturation tests indicated poor phylogenetic signal at this locus, and transitions appeared to be heavily saturated on a plot of substitution type against HKY-corrected genetic distances. It should be noted that although mtCOI has already proven to be a poor locus for phylogenetic reconstruction for this genus, it has better indicators under both of these tests than the LTRS locus did. Heuristic MP search of complete data, as well as transition-only and transversion-only partitions, generated a pattern suggesting extreme homoplasy among individuals of the outgroup taxon Notochthamalus scabrosus, with individuals scattered throughout basal nodes and branches of the tree. These individuals were removed from subsequent analysis at this locus, as the data may be derived from a paralogous gene region or other artifact. Extreme conflicts (e.g., multiple distant clades of the same nominal species) were also identified for C. malayensis and C. neglectus, for which there were few sequences to begin with, so these taxa were also excluded. Finally, n = 8 individuals from the TEP (primarily C. southwardorum and C. panamensis) deviated strongly in phylogenetic placement from nominal conspecifics. Because in general the TEP species tended to each be reciprocally monophyletic with similar topology as in other gene regions, and at least 10 individuals from each of these taxa remained after the 8 sequences are removed, we excluded the deviant sequences from further analysis at this locus. From the subtracted data set, the average internal node bootstrap support was 59.7%, with individual species and internal node support substantially lower using transversions only (similar results using NJ). The best-fit model at this locus was HKY with gamma-distributed site rates (α = 0.39).
At the nuclear NAKAS gene, we obtained 274 bp of aligned sequence data for 109 individual barnacle specimens (Genbank FJ862666-774). Of these characters, 185 were constant and 12 were uninformative; 77 characters were parsimony informative. Saturation tests indicated little-to-moderate saturation at this locus, and a plot of transitions and transversions against corrected genetic distances suggested that this locus behaves well for phylogenetic analysis. Bootstrap support for the unweighted MP tree suggests expected species-group support (e.g., C. dalli and C. challengeri grouped with 97% bootstrap support) but very little resolution at deeper nodes of the tree (similar results with NJ). The best-fit model of molecular evolution at this locus was F81 (substitution rates equal) with invariant sites (p = 0.47) and gamma-distributed rates (α = 0.65). Base frequencies were G/C biased (62%); both models were compared for independent analyses at this locus.
At the nuclear EF1a locus, we obtained 464 bp of aligned sequence data for 196 individual specimens (Genbank FJ862307-502). Of these data, 319 characters were constant and 119 were parsimony informative, with no gaps. Saturation tests indicated little saturation at this locus, and a plot of substitution types against HKY-corrected distances suggested this locus could behave well for phylogenetic anlaysis. Most species delineated by mtCOI were well supported by EF1a; C. dalli was only separable from C. challengeri in 57% of bootstrap replicates, and the cryptic group of C. angustitergum was separable 59% of the time. The TEP group was again poorly separated. Deeper nodes were not well supported in the MP bootstrap analysis, with average internal node bootstrap support only 52%; results were comparable for NJ bootstrap. The best-fit model of molecular evolution at this locus was TrN with invariant sites (p = 0.52) and gamma rates (α = 0.97).
We obtained 817 bp of aligned sequence data at the nuclear 18S rDNA locus for 20 specimens across 20 species. Of these data, only 9 characters were parsimony informative. Saturation testing indicated little saturation; however, this locus harbored too little variation to reliably distinguish even the outgroup taxa from our ingroup (although the Microeuraphia species grouped together with 84% support, no other species pair had higher than 67% support and the tree remained otherwise completely unresolved). Since there was so little variation among even distantly related taxa at this locus, it was excluded from subsequent analysis.
For the concatenated data set, we obtained a complete data set of 1671 characters for 145 taxa. Of these data, 558 characters were parsimony-informative. Bootstrap analysis was consistent with single-gene approaches in the strong support for individual taxa. Excluding the bootstrap support for these tip clades, the average bootstrap support for internal nodes was 79%, ranging from ≤50 to 100%. Similar results were obtained using NJ and Bayesian analysis.
Following these preliminary analyses, two cryptic clades of apparent Microeuraphia species, both sampled in the TEP, were excluded because of inconsistencies among gene partitions and inconsistent placement among topologies chosen within gene partitions and concatenated data sets. One of these clades was related to the Jalisco specimens of Wares  and no morphological specimens remained; the other clade was represented by only 3 individuals that suggested a reciprocally monophyletic clade at mtCOI, but placement of these individuals varied widely among other genes. Final analysis proceeded with the reduced data set for full Bayesian analysis, with support under other criteria (parsimony, distance) indicated where appropriate.
Full Phylogenetic Analysis of Chthamalus
Given the final data set, re-analysis of each gene proceeded individually and with the concatenated complete data. Results for individual genes are shown in Figure 1. As suggested by analysis of the unfiltered data, the nuclear genes generated less consistent support for some individual taxa and most of the deeper nodes of the phylogeny. Figure 1A shows an almost-fully resolved phylogeny at mt16S, with a mean posterior probability at internal nodes of 0.879. The ‘fissus group’ was monophyletic and completely supported, as was the ‘challengeri group+Microeuraphia’ clade; however, the tree was not resolved for the relationship of these two clades, C. bisinuatus, and the ‘malayensis group’. The three nuclear genes (Figure 1B–D) provided strong support for some species-level differentiation, but internal nodes were lacking in support. This was consistent with low mean internal node bootstrap support at these loci (e.g., 59.7% at nLTRS as shown above). The nLTRS locus exhibited a homoplasy index in parsimony analysis of 0.40, with a rescaled consistency index of only 0.56. Similar results were obtained at all 3 nuclear loci.
Phylograms generated from individual Bayesian analysis of (A) mt16S, (B) nLTRS, (C) nNAKAS, and (D) nEF1 gene regions. Nodes with less than 0.50 posterior probability collapsed. Illustrated as “circle trees” to provide spatial separation at small scale (and for cirripedic appearance). Posterior support for nodes illustrated via thickness of branches; posterior probabilities of nodes shown in (A) mt16S except where posterior = 1.0, same thickness scale used on trees (B–D). Taxa that were found to be wholly monophyletic (based on categorization with mtCOI) are indicated as cartoons with depth of intraspecific variation portrayed as depth of triangle (using FigTree v.2.1). At (A) mt16S, all species are indicated, with asterisks indicating TEP species for which a single individual appeared in the incorrect but otherwise monophyletic grouping of that taxon. Species abbreviated with first letter of genus and first two letters of specific epithet, except for C. angustitergum = C. ag. As with parsimony and neighbor-joining analysis, there is little deep-node resolution at the nuclear gene regions. In (A) mt16S, outgroup taxa are indicated in grey; ‘challengeri group’+Microeuraphia indicated in red (posterior probability 0.83); ‘fissus group’ indicated in blue (posterior probability 1.0).
Concatenated analysis of sequence data generated a more reliable tree. Including all 4 gene regions allowed reconstruction of sister taxon relationships with high confidence (e.g., C. stellatus/C. montagui, C. challengeri/C. dalli, C. angustitergum/C. bisinuatus, C. fissus/C. anisopoma, C. stellatus/C. fragilis all supported with 100% posterior probability), but the backbone of the tree had almost no statistical support. The ‘fissus group’ was poorly resolved and had only 0.50 posterior probability as a clade, for example, and overall the internal node mean posterior support was only 0.73 (including the branches supporting sister taxa; without these branches mean posterior support was only 0.60). Subsequent analysis excluded each of the four primary loci individually to generate a concatenated 3-locus phylogeny; significant improvement was obtained through exclusion of the nLTRS locus. This approach was confirmed through pairwise partition homogeneity tests of our loci; the nLTRS locus, in all pairwise comparisons, appeared to be significantly incompatible (p<0.001). Only 1 of 3 remaining pairwise tests (mt16S vs. nEF1a) generated a similar result, but each of these loci were easily compatible with nNAKAS (p>0.025 for both). Nearly identical results were obtained when invariant characters were excluded, as shown in simulation studies ; since significance probabilities for this test tend to indicate improved phylogenetic signal when p>0.01, and the test in general is overly conservative , , nEF1a was retained.
Figure 2 illustrates this ‘final’ phylogenetic analysis, which provides evidence for a wholly monophyletic and strongly supported ‘fissus group’, and high statistical support throughout the tree with the exception of a statistical polytomy among the ‘challengeri group+Microeuraphia’, ‘C. stellatus+C. montagui’, and ‘fissus group’ clades. These results were reflected by poorly correlated (r = 0.32, p>0.15) parsimony bootstrap support; many deep nodes have poor support under parsimony that were well reconstructed by the independent-model likelihood approach.
Phylogram generated from Bayesian analysis of concatenated mt16S, nNAKAS, and nEF1 gene regions with unlinked substitutional models. For each clade, the posterior probability is indicated (with numerical support and correlated branch thickness/color), followed by the unweighted parsimony nonparametric bootstrap support at the same node (separated by ‘/’). Correlation of Bayesian and parsimony support indices was only 0.44 at internal nodes, with similar support from neighbor-joining bootstrap analysis. The “fissus group”, indicated with species names in bold, is monophyletic and strongly supported under all 3 methods (posterior probability 1.0, MP bootstrap 99%, NJ bootstrap 90%).
Bayesian Estimate of Species Tree
While this analytical approach was used to attempt resolution of the most difficult nodes on our reconstructed phylogeny through using intraspecific variation as a cue to temporal order of species divergence, we were unable to reconstruct a satisfactory species tree after repeated attempts at parameter optimization. In every individual mb-best analysis, the duplicate runs did not converge, with average standard deviation of split frequencies larger than 0.10. This result was obtained whether using the whole concatenated data set or excluding LTRS as above. Final analytical results approximated the species tree shown in Figure 2, but with very low differentiation among possible topologies in terms of their posterior probability. No single species tree topology carried a posterior probability >0.016, with the top 19 topologies (out of 2963 recovered) accounting for the first 10% of the cumulative posterior distribution. The subgroup clades indicated in Figure 2 were not wholly monophyletic with respect to prior taxonomy, e.g. C. dentatus fell out with the basal taxa C. neglectus and C. malayensis in some of these ‘high probability’ trees, although the general relationships in Figure 2 that are of note here – the monophyly of the ‘fissus group’ (other than C. dentatus, in this case) and the recent relationship between species in the ‘challengeri group’ and Microeuraphia – were supported in the final consensus tree and the most probable ultimate topologies.
Species Radiation Timing
The divergence of tropical barnacles with the rise of the Panamanian Isthmus not only generates diversity at the species level, it also allows calibration of the rate of molecular evolution for novel taxa and, in this case, novel genes. Table 1 illustrates the distribution of rates associated with comparisons of the two geminate species of Microeuraphia and the mean divergence (net nucleotide distance) among trans-Isthmian Chthamalus species (i.e. C. proteus+C. fragilis vs. C. fissus+C. anisopoma+the TEP species). Phylogenetic reconstruction at each gene region deviated significantly from the expectations of a molecular clock (evaluated by standard likelihood ratio test). Additional corrections to this model involve additional assumptions and may still be misleading , and no precise temporal hypotheses are being evaluated with these data, therefore no additional testing or adjustment of rates was done. Overall, the observed rate at mt16S (about 1.25% per million years) was comparable to previous studies at this locus, while the first-time rate estimates at the other three loci appeared to be about half as fast (Table 1). At the nLTRS locus, substantial differentiation between M. rhizophorae and M. eastropacensis (see Figure 1B) suggested a much higher rate than estimation based on divergence in Chthamalus alone. Given the unusual patterns of substitution noted in other taxa (e.g. N. scabrosus) this gene region may violate assumptions of rate constancy.
Knowing that several speciation events have occurred in Chthamalus after the closure of the Panamanian Isthmus (C. proteus−C. fragilis, and the divergence between the TEP species and the C. fissus−C. anisopoma pair, and divergences within each of those clades) suggests high levels of speciation in the Pleistocene. A character-independent analysis using the BiSSE model was not able, however, to distinguish the gene tree for Chthamalus from one in which speciation rate is equivalent to extinction rate (i.e. a Yule uniform speciation model, results not shown). Coding the species in the ‘fissus group’ separately from other species in the phylogeny (to examine the diversification of this group in the Americas using the character-associated diversification BiSSE analysis) identified slightly elevated speciation rate λ for this group (4.104 versus 3.859), but the extinction rate μ was considerably lower (7.826×10−6 versus 2.524). However, a complete phylogeny of the remaining groups will be necessary to fully evaluate this pattern.
Two consistent results are gained from this extensive phylogenetic analysis of Chthamalid barnacles: first, a dominant upper intertidal group in North America and northern South America, the ‘fissus group’ sensu Dando and Southward , is monophyletic. The internal morphological characters (bidenticulate setae with basal guards on cirrus II, absence of conical spines on cirrus I) are not exclusive to this group, and confirmation with molecular data is important for understanding the radiation of this genus along both coasts of the Americas. Pitombo (in prep) has identified a novel synapomorphy for this subgeneric group, and the monophyly of all four groups (based on a combination of presence/absence of the characters above) still needs to be evaluated (see below).
The second result, and one that is of greater phylogenetic interest, is that Microeuraphia appears not to be the ancestral or sister taxon to Chthamalus as is frequently suggested in the literature , , but is consistently nested within Chthamalus and is closely affiliated with species in the current ‘challengeri group’. These results are supported by an independent analysis (Y. Achituv, unpublished) and are consistent across various phylogenetic models and reconstruction criteria. Here we use three outgroup genera, two (Semibalanus and Amphibalanus) from families Archaeobalanidae and Balanidae (Leach 1817) and all other species in this study from the family Chthamalidae (Darwin 1854), and obtain strong bootstrap and posterior probability for the inclusion of Microeuraphia – originally presumed to be a fourth outgroup taxon – within the Chthamalus clade. Given the uncertainty at some internal nodes on this tree, we additionally examined the likelihood that Chthamalus is monophyletic with all other genera forming a reciprocal clade by performing a Kishino-Hasegawa test to compare unconstrained trees vs. trees constrained to include Semibalanus, Amphibalanus, Notochthamalus, and Microeuraphia in the outgroup. This test indicated a significant (p<0.001) difference between these trees, with a mean difference in log L of 62.54, supporting the monophyly of Chthamalus (including Microeuraphia). Of course, only two species of Microeuraphia are included in the present study, but two other cryptic lineages present in preliminary analyses (identified at the level of mtCOI, but with inconsistent results in other loci) indicated the same pattern, and other studies (Y. Achituv, unpublished) also show Microeuraphia nested within Chthamalus. We also have specimens of M. intertexta and M. aestuarii but did not obtain sufficient data from these species to include; further work can confirm this pattern, but we see no reason to doubt this placement and hence recommend taxonomic revision. Our results reflect some uncertainties already indicated about the relationships of the subfamilies and genera of the Chthamalidae .
Wares had used Microeuraphia as the sole outgroup taxon for rooting a mitochondrial phylogeny of Chthamalus. This may have led to a variety of inconsistencies that are resolved in this study by more extensive genomic and species sampling as well as more detailed curation of morphological specimens. The phylogeny shown in Figure 2 is effectively only a phylogeny of the consistent, with regard to criterion (the same topology obtained using NJ, MP, and Bayesian/likelihood approaches), gene region, and OTU. Specimens that generated more complex patterns – including those noted in results such as the LTRS data in Notochthamalus and some TEP specimens that were found in different clades depending on the gene region analyzed – are not presented in the final analysis. There is likely more phylogenetic and biological complexity to be represented due to potential hybridization among taxa and other forms of intraspecific variation that are not yet accounted for. Since some of these gene regions are novel for phylogenetic analysis, we may also need to learn more about the mutational models that generate both high diversity and high homoplasy, as in the nLTRS gene region. The concatenated data set including the nLTRS locus had significantly poorer statistical resolution than the analysis when this locus is excluded, despite the large number of informative characters. High levels of substitutional saturation are likely the problem, suggesting that even in this intron region there may be areas of higher mutational frequency.
Generating genetic data that are sufficiently variable for phylogenetic analysis of species that appear to have largely radiated in the Miocene is challenging. As such, we recovered loci that not only differentiate the species well (though not always with high statistical support, Figure 1) but also harbor intraspecific variation. These data should have been appropriate for Bayesian estimation of the species tree , incorporating the concept of the ‘deep coalescent’  in the reconstruction of relationships among shallow-water barnacles. The probability of allelic diversity in ancestral populations – such as the Miocene progenitor population to the challengeri, stellatus, and fissus groups (and Microeuraphia) – sorting locus by locus in a way that does not necessarily follow the containing species tree is by now a well known problem . Other attempts to use this new method for estimating the species tree from multi-allelic information within each taxon have provided results more consistent with expected and observed phylogeny using more traditional methods , though the mb-best approach often indicated much weaker node support across the phylogeny as with our results. Their analysis was primarily focused on intraspecific variation and could apply to the TEP taxa; in all mb-best analyses we did recover a monophyletic TEP clade within the ‘fissus group’, for example. Given the exceedingly low posterior probabilities for most nodes found in our mb-best analysis (as well as in ), along with topologies that largely disagree with traditional phylogenetic approaches (Figure 2 and methods above) as well as traditional taxonomy, we assume that the radiation of these species was simply very rapid at certain points in their history (i.e. associated with nodes around the noted region in Figure 2), leading to extensive hemiplasy , and that other approaches for inferring the importance of deep coalescent events  in our data may prove useful.
The rapid radiation of species certainly applies to the morphologically and ecologically cryptic TEP species (C. southwardorum, C. hedgecocki, and C. panamensis). Given their position in the topology and our understanding of trans-Isthmian relationships, high level of speciation must have taken place in the Pleistocene in the tropical eastern Pacific. It may be questioned how real these species are under the conditions of DNA barcoding: while there are 4 reciprocally monophyletic taxa under analysis of mtCOI (Appendix I), in no case do these groups fit the “10× rule”  that might indicate substantial temporal and demographic isolation that may be predictive of reproductive isolation. However, the taxa are also recovered using 16S as distinguishable clades, though not reciprocally monophyletic (i.e., C. southwardorum B, the northern clade, appears to be nested within the southern group of C. southwardorum A). They are statistically distinguishable but not monophyletic groups at the nuclear loci (Meyers, Pankey, and Wares in prep). While there are clear phylogeographic patterns resolved in C. southwardorum (and also C. angustitergum on the Caribbean coast), it is important to recognize that these patterns may require substantial effort to resolve further. Dando and Southward  illustrate the potential for large disjunction in available habitat for C. angustitergum on the Central American coast suggesting that while the divergence between populations of this species in Panama and Belize may be significant (reciprocally monophyletic at mtCOI and divergent at mt16S and other nuclear loci in this paper), there may be little more to do in describing this disjunction - perhaps an incipient speciation - if few populations exist in between these areas. The divergence between northern and southern populations of C. southwardorum is also important and perhaps maintained by the lack of available habitat across southwestern Mexico and northern Central America , but given the morphological similarity and broad geographic overlap with other species  in the Tropical Eastern Pacific, a phylogeographic survey of this species would probably require significant additional sampling of congeners, but would perhaps help identify the microhabitat that is more predictive of the presence of C. southwardorum .
An important element of phylogenetic analysis in a widely distributed marine genus such as Chthamalus is the ability to expand on the traditional sequence markers that have been used for biogeographic or phylogenetic study. With two sets of trans-Isthmian comparisons to make for all loci considered here, we are able to identify likely evolutionary rates for at least two novel markers that may be useful in subsequent phylogeographic and phylogenetic analysis (Table 1). Outstanding issues that remain to be explored include a more detailed phylogeny incorporating the eastern hemisphere taxa, particularly the “malayensis group” (for which additional specimens were available but difficult to curate with respect to recognized intraspecific variation; COI data from C. malayensis in this study correspond with the ‘South China Sea’ clade from Tsang et al. , but other specimens generated unrecognizable sequence diversity). Pope  and Southward and Newman  noted that additional investigation is necessary for both the malayensis and challengeri subgroups, and inclusion of only 2 taxa per group in this study is sufficient only to show that they are likely to be monophyletic groups. However it is also clear that for taxa that have radiated as recently as the genus Chthamalus, the search for adequate phylogenetic information is still incomplete. Comparison of genomic data sets from related taxa may improve our ability to identify gene regions with sufficient variability and consistency for full resolution of this phylogenetic problem .
Overall, the results of our study point to a pattern that suggests rapid radiation in this genus, particularly along the American coasts, in the past few million years. Further analysis of the TEP species of Chthamalus – particularly with regard to the separation of C. southwardorum clades across the Central American Faunal Gap  – will improve our understanding of the association between species radiation in this group and specific climatic and oceanographic events that may have affected the paleodistributions and interactions of these species . Although little is known about the microhabitat requirements for codistributed species of Chthamalus , , , there is an intriguing frequency of closely related species in sympatry in the TEP (Hellberg 1998), which suggests that simultaneous evaluation of speciation and species interactions in this group will be fruitful. We have also recovered, using advanced phylogenetic and deep coalescent approaches, a relationship between the genera Chthamalus and Microeuraphia that not only suggests taxonomic revision is necessary, but a re-evaluation of our understanding of the morphological evolution of this family of barnacles . The barnacles in the genus Chthamalus are small and relatively nondescript, but these high intertidal species have contributed to seminal work in ecology and biogeography, and a complete understanding of the evolutionary history of this group enables a more thorough understanding of the causes and consequences of macroevolutionary patterns in barnacles globally .
Materials and Methods
Collection and Curation of Specimens
Table 2 lists all sites from which specimens were obtained under permit, along with sample sizes, identification, and GenBank accession information. From each individual, the musculature connecting soma to opercular plates was dissected away, leaving key morphological features (e.g., cirri, trophi) for taxonomic verification. Remaining taxonomically relevant tissues are preserved in 95% ethanol for further examination. Nucleic acids were isolated from musculature using the PureGene protocol (Gentra Systems), with template DNA quantified on a Nanodrop spectrophotometer.
After coarse morphological identification (cirri I/II, mandible), individual specimens were ‘barcoded’ using universal  mitochondrial cytochrome oxidase I (mtCOI) primers for comparison with previous taxonomic and phylogenetic analyses of this genus (e.g., , , , ). A sample of individuals from each mtCOI clade (see Table 2) identified as having 100% bootstrap support (neighbor-joining algorthim in GENEIOUS 2.5.4 and/or PAUP*4.0b10, Swofford ), 100% consensus support, or 100% posterior Bayesian probability (MrBayes 3.1, ) was then sequenced at additional loci. These barcoded OTUs – that in almost every case match their identity based on morphology and sample location – are used independently of subsequent phylogenetic analysis, because studies ,  have illustrated the poor performance of mtCOI for phylogenetic analysis of this group.
Sequence Collection and Verification
A sample of individuals from each mtCOI clade identified in Table 2 was then sequenced for 5 additional gene regions: mitochondrial 16S rDNA (mt16S F: 5′-CTGTGCTAAGGTAGCATAATCA; R: 5′-AGAAGATAGAAACCAACCTG), nuclear lysidyl aminoacyl transfer RNA synthetase (nLTRS;; F: 5′-CGAATGGATGACACGACGTA; R: 5′-GGATGGGTTCATTTTCAAGG), nuclear eukaryotic elongation factor 1α (nEF1 F: 5′-CAGACGCAGGGGCTTGTC; R: 5′-GCCACAGGGATTTCATCAAG), nuclear r18S , and nuclear Na-K-ATPase (nNAKA F: 5′-GTGGTTCGACAACCAGATCA; R: 5′-GGGATCTCGCAGACCTTCTT). These loci were chosen based on previous phylogenetic performance in closely related taxa (16S, 18S, and nEF1; ) or were developed from previously published regions that harbor high variability in crustaceans (LTRS, , NAKAS, ). The newly-developed loci represent non-coding intron data (LTRS) and coding sequence (NAKAS) that harbors substitutions consistent with neutral evolution in interspecific comparisons (McDonald-Kreitman test, p>0.05, Wares, Zakas, and Pankey unpublished results). DNA from 3–10 specimens per species (per location for suspected cryptic species) was analyzed.
PCRs were carried out in 25 µl volumes consisting of 0.5 µM each primer, 0.8 mM total dNTPs, 2 mM MgCl2, and 1 U Taq polymerase (Promega). Annealing temperatures for each locuswere as follows: 16S, 50°; COI, 48°; EF1, 52°; LTRS, 47°; NAKA, 46°; r18S, 50°. PCR products were prepared for sequencing using Exonuclease I and Antarctic Phosphotase (NEB). Sequencing reactions were carried out in 10 µl volumes with 80 ng of prepared template, 0.6 µM primer, 0.6 µl BigDye Terminator (Applied Biosystems) and 3.4 µl Better Buffer (The Gel Company). Sequences were cleaned with 4 volumes 75% isopropanol, suspended in Hi-Di formamide (Applied) and run on an ABI 3730.
For each locus, sequence data were edited using CodonCode Aligner v.2.06. All nucleotide data were scored for quality using PHRED , with sites having quality <15 declared ambiguous (N). Sequences included for analysis were end-clipped so that only bases with quality scores of 15 and higher were retained. Sequences were aligned using Aligner's built-in ‘end-to-end’ algorithm (with mismatch and gap penalties of −2 and additional gap penalty of −3; other alignment parameters were evaluated with little change in alignment quality), examined and edited for likely artifacts caused by poly-N repeats and other apparent insertions, and disassembled/realigned. In all data sets, gapped characters are removed from analysis. Each data set was then examined for phylogenetic signal using DAMBE  and the method for inferring substitution saturation developed by Xia and colleagues .
Each sequence data set was first subjected to maximum parsimony (MP) and uncorrected neighbor-joining (NJ) analysis in PAUP*4.0b10  with 1000 nonparametric bootstrap replicates, using tree-bisection-reconnection rearrangements and a heuristic search, holding the maximum number of equal trees at 100. These preliminary analyses evaluated two components necessary for subsequent species-level analysis. First, we identified sequences that do not appear to consistently group with apparent barcoded conspecifics; these sequences may represent curation errors, artifacts, or chimeric sequences generated naturally. Such sequences were excluded from subsequent analysis (see Results). Additionally, minimum and mean bootstrap values were identified for each locus. There is evidence that these measures are indicators of the likelihood that a data set can recover the true phylogeny . The phylogenetic signal at each locus was evaluated in pairwise comparisons with each other locus using a partition homogeneity test  implemented in PAUP*. Each comparison involved 1000 permutations and was repeated with and without constant characters. A concatenated data set (generated with the fused matrix export option in Mesquite and using only individuals for which at least 2 loci were fully sequenced, with missing information coded distinctly from gapped characters) was evaluated for similar performance.
The best-fit model of molecular evolution was assessed for each locus using Modeltest 3.7 , using likelihood-ratio tests to choose the simplest model with a significantly improved fit to the data. For subsequent Bayesian analysis, we used MrModeltest  instead; similar models were generated for each locus by both model estimation procedures. These models are used independently for each gene region for all subsequent distance, likelihood, or Bayesian phylogenetic analyses discussed below. Distance, parsimony, and likelihood analysis of each data set was performed in PAUP*4.0b10 with random addition, tree-bisection-reconnection branch swapping, and heuristic search methods. For Bayesian analysis, each locus was analyzed separately using MrBayes, with model parameters as chosen above, 4 heated MCMC chains and 2 independent runs in each analysis. Bayesian analysis involved a minimum of 5×106 MCMC generations with sample frequency f = 1000; if the standard error between independent runs was not less than 0.05 at this point, additional sets of 1×106 generations were run until this standard was reached. Unlinked concatenated gene analysis on a reduced data set (n = 3+ individuals per species) was analyzed in the same way, as in . The first 25% of generations were discarded as burnin, an approach verified through graphical analysis of the stabilization of likelihood values.
Bayesian Estimate of Species Tree
Subsequent to traditional phylogenetic analyses, a multi-locus approach that allowed for distinct genealogical histories within the species tree (and used the variation among gene trees as information regarding the interaction of ancestral effective population size and time between nodes on ability to resolve those nodes) was attempted (see  for review). This approach uses mb-best v. 2.0 , a modification of MrBayes. As before, this Bayesian analysis used unlinked models of molecular evolution for each locus. Four initial attempts were used to generate the best-performing set of prior distributions on per-gene theta, with each run requiring more than 30 days of computational time on 2.0 GHz dual-processor Macintosh computers. Runs utilized 2×107 generations with four MCMC chains and a sample frequency f = 1000 trees. We chose an inverse gamma with parameters [2,1] for theta in our final analyses, with a default uniform prior [0.5,1.5] on gene mutation rate.
Species Radiation Timing
To evaluate the rate of molecular evolution at each locus, standard likelihood ratio tests were performed on individual loci as in , utilizing the best-fit model and the same model constrained by clock assumptions. Following these tests, trans-Isthmian comparisons were made of the geminate pair of Microeuraphia species, and for all pairs of trans-Isthmian Chthamalus species; the latter arrangement is based on speciation that has taken place in the Caribbean (C. proteus/C.fragilis) and Pacific (C. fissus, C. anisopoma, and the TEP species of ) subsequent to closure of the Panamanian Isthmus. Hence, the net nucleotide divergence for all species pairs was calculated in DNAsp v.4.20  and averaged for the Chthamalus comparison, with standard deviation estimated and summarized the same way. Rate estimates are based on the range of closure/isolation dates from 3.0–3.5 mya .
The final concatenated topology was evaluated using analysis of speciation and extinction in the BiSSE character-independent model . This analysis generates relative rates of speciation and extinction through time, and calculates the likelihood that one rate is higher than the other for a given phylogeny with branch lengths. This analysis was repeated using the character-associated BiSSE model to examine different levels of diversification rates between the mostly American ‘fissus group’ and remaining taxa on the phylogeny.
This work is dedicated to Alan Southward (1928–2007) and Bill Newman, who helped many of us along in our study of barnacles. Significant administrative, technical, and field assistance was provided by M. Frey, R. Miller, T. Bell, S. Small, C. Zakas, J. Robinson, J. Binford, K. Stanger-Hall, M. Hellberg, M. Hendricx, J. Zardus, K. Vicknair, E. Sotka, S. Navarrete, N. Blecher, K. Crandall, S. Jenkins, C. Embach, E. Díaz-Ferguson, and the Smithsonian Tropical Research Institute.
Conceived and designed the experiments: JW YA. Performed the experiments: JW MSP. Analyzed the data: JW MSP. Contributed reagents/materials/analysis tools: JW FP LGD YA. Wrote the paper: JW MSP.
- 1. Darwin C (1854) A monograph on the sub-class Cirripedia. The Balanidae, etc. London: The Ray Society. C. Darwin1854A monograph on the sub-class Cirripedia. The Balanidae, etcLondonThe Ray Societyviii + 684
- 2. Connell JH (1961) The influence of interspecific competition and other factors on the distribution of the barnacle Chthamalus stellatus. Ecology 42: 710–723.JH Connell1961The influence of interspecific competition and other factors on the distribution of the barnacle Chthamalus stellatus.Ecology42710723
- 3. Burrows MT, Hawkins SJ, Southward AJ (1992) A comparison of reproduction in co-occurring chthamalid barnacles, Chthamalus stellatus (Poli) and Chthamalus montagui Southward. J Exp Mar Biol Ecol 160: 229–249.MT BurrowsSJ HawkinsAJ Southward1992A comparison of reproduction in co-occurring chthamalid barnacles, Chthamalus stellatus (Poli) and Chthamalus montagui Southward.J Exp Mar Biol Ecol160229249
- 4. Wethey DS (2002) Biogeography, competition, and microclimate: the barnacle Chthamalus fragilis in New England. Integrative and Comparative Biology 42: 872–880.DS Wethey2002Biogeography, competition, and microclimate: the barnacle Chthamalus fragilis in New England.Integrative and Comparative Biology42872880
- 5. Southward AJ (1991) Forty years of changes in species composition and population density of barnacles on a rocky shore near Plymouth, England, UK. J Mar biol Ass UK 71: 495–514.AJ Southward1991Forty years of changes in species composition and population density of barnacles on a rocky shore near Plymouth, England, UK.J Mar biol Ass UK71495514
- 6. Barry JP, Baxter CH, Sagarin RH, Gilman SE (1995) Climate-related, long-term faunal changes in a California rocky intertidal community. Science 267: 672–675.JP BarryCH BaxterRH SagarinSE Gilman1995Climate-related, long-term faunal changes in a California rocky intertidal community.Science267672675
- 7. Pitombo FB, Burton RS (2007) Systematics and biogeography of Tropical Eastern Pacific Chthamalus with descriptions of two new species (Cirripedia, Thoracica). Zootaxa 1574: 1–30.FB PitomboRS Burton2007Systematics and biogeography of Tropical Eastern Pacific Chthamalus with descriptions of two new species (Cirripedia, Thoracica).Zootaxa1574130
- 8. Yan Y, Chan BKK (2004) Larval morphology of a recently recognized barnacle, Chthamalus neglectus (Cirripedia: Thoracica: Chthamalidae), from Hong Kong. J Crust Biol 24: 519–528.Y. YanBKK Chan2004Larval morphology of a recently recognized barnacle, Chthamalus neglectus (Cirripedia: Thoracica: Chthamalidae), from Hong Kong.J Crust Biol24519528
- 9. Dando PR (1987) Biochemical genetics of barnacles and their taxonomy. In: Southward AJ, editor. Rotterdam: A.A. Balkema. pp. 73–88.PR Dando1987Biochemical genetics of barnacles and their taxonomy.AJ SouthwardRotterdamA.A. Balkema7388
- 10. Pfenninger M, Schwenk K (2007) Cryptic animal species are homogeneously distributed among taxa and biogeographical regions. BMC Evolutionary Biology 7: 121.M. PfenningerK. Schwenk2007Cryptic animal species are homogeneously distributed among taxa and biogeographical regions.BMC Evolutionary Biology7121
- 11. Poltarukha OP (1997) Composition, phylogeny, and position of the subfamily Euraphiinae (Crustacea, Chthamalinae) in the system of Cirripedia. Russian Journal of Zoology 1: 463–470.OP Poltarukha1997Composition, phylogeny, and position of the subfamily Euraphiinae (Crustacea, Chthamalinae) in the system of Cirripedia.Russian Journal of Zoology1463470
- 12. Pilsbry HA (1916) The sessile barnacles (Cirripedia) contained in the collections of the United States National Museum; including a monograph of the American species. United States National Museum Bulletin 93: 1–366.HA Pilsbry1916The sessile barnacles (Cirripedia) contained in the collections of the United States National Museum; including a monograph of the American species.United States National Museum Bulletin931366
- 13. Nilsson-Cantell CA (1921) Cirripeden Studien Zur Kenntnis der Biologie, Anatomie, und Systematick dieser Gruppe. Zool Bidr Uppsala 7: 75–390.CA Nilsson-Cantell1921Cirripeden Studien Zur Kenntnis der Biologie, Anatomie, und Systematick dieser Gruppe.Zool Bidr Uppsala775390
- 14. Zullo VA (1963) A classification and phylogeny of the Chthamalidae (Cirripedia: Thoracica). XVI International Congress of Zoology 1: 190.VA Zullo1963A classification and phylogeny of the Chthamalidae (Cirripedia: Thoracica).XVI International Congress of Zoology1190
- 15. Pope EC (1965) A review of Australian and some Indomalayan Chthamalidae (Crustacea: Cirripedia). Proc Linn Soc New South Wales 90: 10–77.EC Pope1965A review of Australian and some Indomalayan Chthamalidae (Crustacea: Cirripedia).Proc Linn Soc New South Wales901077
- 16. Dando PR, Southward AJ (1980) A new species of Chthamalus (Crustacea: Cirripedia) characterized by enzyme electrophoresis and shell morphology: with a revision of other species of Chthamalus from the western shores of the Atlantic Ocean. J Mar Biol Assoc UK 60: 787–831.PR DandoAJ Southward1980A new species of Chthamalus (Crustacea: Cirripedia) characterized by enzyme electrophoresis and shell morphology: with a revision of other species of Chthamalus from the western shores of the Atlantic Ocean.J Mar Biol Assoc UK60787831
- 17. Southward AJ, Newman WA (2003) A review of some common Indo-Malayan and western Pacific species of Chthamalus barnacles (Crustacea: Cirripedia). J Mar biol Ass UK 83: 797–812.AJ SouthwardWA Newman2003A review of some common Indo-Malayan and western Pacific species of Chthamalus barnacles (Crustacea: Cirripedia).J Mar biol Ass UK83797812
- 18. Newman WA, Ross A (1976) Revision of the balanomorph barnacles; including a catalog of the species. San Diego Society of Natural History Memoirs 9: 1–108.WA NewmanA. Ross1976Revision of the balanomorph barnacles; including a catalog of the species.San Diego Society of Natural History Memoirs91108
- 19. Southward AJ, Newman WA (1977) Aspects of the ecology and biogeography of the intertidal and shallow-water Balanomorph Cirripedia of the Caribbean and adjacent sea-areas. FAO Fisheries Report 200: 407–425.AJ SouthwardWA Newman1977Aspects of the ecology and biogeography of the intertidal and shallow-water Balanomorph Cirripedia of the Caribbean and adjacent sea-areas.FAO Fisheries Report200407425
- 20. Hedgecock D (1979) Biochemical genetic variation and evidence of speciation in Chthamalus barnacles of the tropical eastern Pacific ocean. Marine Biology 54: 207–214.D. Hedgecock1979Biochemical genetic variation and evidence of speciation in Chthamalus barnacles of the tropical eastern Pacific ocean.Marine Biology54207214
- 21. Wares JP (2001) Patterns of speciation inferred from mitochondrial DNA in North American Chthamalus (Cirripedia: Balanomorpha: Chthamaloidea). Molecular Phylogenetics and Evolution 18: 104–116.JP Wares2001Patterns of speciation inferred from mitochondrial DNA in North American Chthamalus (Cirripedia: Balanomorpha: Chthamaloidea).Molecular Phylogenetics and Evolution18104116
- 22. Fisher T, Katcoff DJ, Achituv Y (2004) A phylogenetic study of chthamaloids (Cirripedia; Thoracica; Chthamaloidae) based on 16S rDNA and COI sequence analysis. Biol J Linn Soc 83: 39–45.T. FisherDJ KatcoffY. Achituv2004A phylogenetic study of chthamaloids (Cirripedia; Thoracica; Chthamaloidae) based on 16S rDNA and COI sequence analysis.Biol J Linn Soc833945
- 23. Peters JL, Zhuravlev Y, Fefelov I, Logie A, Omland KE (2007) NUCLEAR LOCI AND COALESCENT METHODS SUPPORT ANCIENT HYBRIDIZATION AS CAUSE OF MITOCHONDRIAL PARAPHYLY BETWEEN GADWALL AND FALCATED DUCK (ANAS SPP.). Evolution 61: 1992–2006.JL PetersY. ZhuravlevI. FefelovA. LogieKE Omland2007NUCLEAR LOCI AND COALESCENT METHODS SUPPORT ANCIENT HYBRIDIZATION AS CAUSE OF MITOCHONDRIAL PARAPHYLY BETWEEN GADWALL AND FALCATED DUCK (ANAS SPP.).Evolution6119922006
- 24. Tsang LM, Chan BKK, Wu TH, Ng WC, Chatterjee T, et al. (2008) Population differentiation in the barnacle Chthamalus malayensis: postglacial colonization and recent connectivity across the Pacific and Indian Oceans. Mar Ecol Prog Ser 364: 107–118.LM TsangBKK ChanTH WuWC NgT. Chatterjee2008Population differentiation in the barnacle Chthamalus malayensis: postglacial colonization and recent connectivity across the Pacific and Indian Oceans.Mar Ecol Prog Ser364107118
- 25. Degnan JH, Salter LA (2005) Gene tree distributions under the coalescent process. Evolution 59: 24–37.JH DegnanLA Salter2005Gene tree distributions under the coalescent process.Evolution592437
- 26. Newman WA (1987) Evolution of cirripedes and their major groups. In: Southward AJ, editor. Rotterdam: A.A. Balkema. pp. 3–42.WA Newman1987Evolution of cirripedes and their major groups.AJ SouthwardRotterdamA.A. Balkema342
- 27. Darlu P, Lecointre G (2002) When does the incongruence length difference test fail? Mol Biol Evol 19: 432–437.P. DarluG. Lecointre2002When does the incongruence length difference test fail?Mol Biol Evol19432437
- 28. Cunningham CW (1997) Can three incongruence tests predict when data should be combined? Mol Biol Evol 14: 733–740.CW Cunningham1997Can three incongruence tests predict when data should be combined?Mol Biol Evol14733740
- 29. Pringle A, Baker DM, Platt JL, Wares JP, Latge JP, et al. (2005) Cryptic speciation in a global and apparently asexual fungus, the human pathogen Aspergillus fumigatus. Evolution 59: 1886–1899.A. PringleDM BakerJL PlattJP WaresJP Latge2005Cryptic speciation in a global and apparently asexual fungus, the human pathogen Aspergillus fumigatus.Evolution5918861899
- 30. Liu L, Pearl DK, Brumfield RT, Edwards SV (2008) Estimating species trees using multiple-allele DNA sequence data. Evolution 62: 2080–2091.L. LiuDK PearlRT BrumfieldSV Edwards2008Estimating species trees using multiple-allele DNA sequence data.Evolution6220802091
- 31. Avise JC, Robinson TJ (2008) Hemiplasy: A new term in the lexicon of phylogenetics. Syst Biol 57: 503–507.JC AviseTJ Robinson2008Hemiplasy: A new term in the lexicon of phylogenetics.Syst Biol57503507
- 32. Belfiore NM, Liu L, Moritz C (2008) Multilocus phylogenetics of a rapid radiation in the genus Thomomys (Rodentia: Geomyidae). Syst Biol 57: 294–310.NM BelfioreL. LiuC. Moritz2008Multilocus phylogenetics of a rapid radiation in the genus Thomomys (Rodentia: Geomyidae).Syst Biol57294310
- 33. Carling MD, Brumfield RT (2008) Integrating phylogenetic and population genetic analyses of multiple loci to test species divergence hypotheses in Passerina buntings. Genetics 178: 363–377.MD CarlingRT Brumfield2008Integrating phylogenetic and population genetic analyses of multiple loci to test species divergence hypotheses in Passerina buntings.Genetics178363377
- 34. Costa FO, deWaard JR, Boutillier J, Ratnasingham S, Dooh RT, et al. (2007) Biological identifications through DNA barcodes: the case of the Crustacea. Can J Fish Aq Sci 64: 272–295.FO CostaJR deWaardJ. BoutillierS. RatnasinghamRT Dooh2007Biological identifications through DNA barcodes: the case of the Crustacea.Can J Fish Aq Sci64272295
- 35. Hastings PA (2000) Biogeography of the tropical eastern Pacific: distribution and phylogeny of chaenopsid fishes. Zool Journal of the Linn Soc 128: 319–335.PA Hastings2000Biogeography of the tropical eastern Pacific: distribution and phylogeny of chaenopsid fishes.Zool Journal of the Linn Soc128319335
- 36. Wares JP, Castañeda AE (2005) Geographic range in Chthamalus along the west coast of North America. J Mar biol Ass UK 85: 327–331.JP WaresAE Castañeda2005Geographic range in Chthamalus along the west coast of North America.J Mar biol Ass UK85327331
- 37. Kuo C-H, Wares JP, Kissinger J (2008) The Apicomplexan whole genome phylogeny: An analysis of incongruence among gene trees. Mol Biol Evol. C-H KuoJP WaresJ. Kissinger2008The Apicomplexan whole genome phylogeny: An analysis of incongruence among gene trees.Mol Biol Evol
- 38. Briggs JC (1974) Marine zoogeography. New York: McGraw Hill. JC Briggs1974Marine zoogeographyNew YorkMcGraw Hill
- 39. Poloczanska ES, Hawkins SJ, Southward AJ, Burrows MT (2008) Modeling the response of populations of competing species to climate change. Ecology 89: 3138–3149.ES PoloczanskaSJ HawkinsAJ SouthwardMT Burrows2008Modeling the response of populations of competing species to climate change.Ecology8931383149
- 40. Power AM, Piyapattanakorn S, O'Riordan RM, Iyengar A, Myers AA, et al. (1999) Verification of cyprid size as a tool in the identification of two European species of Chthamalus barnacles using mtDNA-RFLP analysis. Marine Ecology Progress Series 191: 251–256.AM PowerS. PiyapattanakornRM O'RiordanA. IyengarAA Myers1999Verification of cyprid size as a tool in the identification of two European species of Chthamalus barnacles using mtDNA-RFLP analysis.Marine Ecology Progress Series191251256
- 41. Stanley SM, Newman WA (1980) Competitive exclusion in evolutionary time: the case of the acorn barnacles. Paleobiology 6: 173–183.SM StanleyWA Newman1980Competitive exclusion in evolutionary time: the case of the acorn barnacles.Paleobiology6173183
- 42. Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology 3: 294–299.O. FolmerM. BlackW. HoehR. LutzR. Vrijenhoek1994DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates.Molecular Marine Biology and Biotechnology3294299
- 43. Zardus JD, Hadfield MG (2005) Multiple origins and incursions of the Atlantic barnacle Chthamalus proteus in the Pacific. Molecular Ecology 14: 3719–3733.JD ZardusMG Hadfield2005Multiple origins and incursions of the Atlantic barnacle Chthamalus proteus in the Pacific.Molecular Ecology1437193733
- 44. Swofford D (2002) D. Swofford2002Phylogenetic Analysis Using Parsimony (PAUP) v. 4.0b10. Phylogenetic Analysis Using Parsimony (PAUP) v. 4.0b10.
- 45. Ronquist F, Huelsenbeck JP (2003) MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574.F. RonquistJP Huelsenbeck2003MRBAYES 3: Bayesian phylogenetic inference under mixed models.Bioinformatics1915721574
- 46. Jarman SN, Ward RD, Elliott NG (2002) Oligonucleotide primers for PCR amplification of coelomate introns. Mar Biotechnol 4: 347–355.SN JarmanRD WardNG Elliott2002Oligonucleotide primers for PCR amplification of coelomate introns.Mar Biotechnol4347355
- 47. Palumbi SR (1996) Nucleic Acids II: The Polymerase Chain Reaction. In: DM H, C M, BK M, editors. Molecular Systematics. 2nd ed. Sunderland: Sinauer Associates, Inc. pp. 205–248.SR Palumbi1996Nucleic Acids II: The Polymerase Chain Reaction.H. DMM. CM. BKMolecular Systematics. 2nd edSunderlandSinauer Associates, Inc205248
- 48. Towle DW, Paulsen RS, Weihrauch D, Kordylewski M, Salvador C, et al. (2001) Na++K+-ATPase in gills of the blue crab Callinectes sapidus: cDNA sequencing and salinity-related expression of alpha-subunit mRNA and protein. J Exp Biol 204: 4005–4012.DW TowleRS PaulsenD. WeihrauchM. KordylewskiC. Salvador2001Na++K+-ATPase in gills of the blue crab Callinectes sapidus: cDNA sequencing and salinity-related expression of alpha-subunit mRNA and protein.J Exp Biol20440054012
- 49. Ewing B, Green P (1998) Base-calling of automated sequencer traces using phred. II. Error probabilities. Gen Res 8: 186–194.B. EwingP. Green1998Base-calling of automated sequencer traces using phred. II. Error probabilities.Gen Res8186194
- 50. Xia X, X Z (2001) DAMBE: data analysis in molecular biology and evolution. J Heredity 92: 371–373.X. XiaZ. X2001DAMBE: data analysis in molecular biology and evolution.J Heredity92371373
- 51. Xia X, Xie Z, Salemi M, Chen L, Wang Y (2003) An index of substitution saturation and its application. Mol Phyl Evol 26: 1–7.X. XiaZ. XieM. SalemiL. ChenY. Wang2003An index of substitution saturation and its application.Mol Phyl Evol2617
- 52. Farris JS, Källersjö M, Kluge AG, Bult C (1994) Testing significance of incongruence. Cladistics 10: 315–319.JS FarrisM. KällersjöAG KlugeC. Bult1994Testing significance of incongruence.Cladistics10315319
- 53. Posada D, Crandall KA (1998) MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817–818.D. PosadaKA Crandall1998MODELTEST: testing the model of DNA substitution.Bioinformatics14817818
- 54. Nylander JAA (2004) MrModeltest v2. Program distributed by the author. JAA Nylander2004MrModeltest v2. Program distributed by the author.Evolutionary Biology Centre, Uppsala University. Evolutionary Biology Centre, Uppsala University.
- 55. Edwards SV (2009) Is a new and general theory of molecular systematics emerging? Evolution 63: 1–19.SV Edwards2009Is a new and general theory of molecular systematics emerging?Evolution63119
- 56. Rozas J, Rozas R (1999) DnaSP version 3: An integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15: 174–175.J. RozasR. Rozas1999DnaSP version 3: An integrated program for molecular population genetics and molecular evolution analysis.Bioinformatics15174175
- 57. Kirby MX, Jones DS, MacFadden BJ (2008) Lower Miocene Stratigraphy along the Panama Canal and Its Bearing on the Central American Peninsula. PLos One 3: e2791.MX KirbyDS JonesBJ MacFadden2008Lower Miocene Stratigraphy along the Panama Canal and Its Bearing on the Central American Peninsula.PLos One3e2791
- 58. Maddison WP, Midford PE, Otto SP (2007) Estimating a binary character's effect on speciation and extinction. Syst Biol 56: 701–710.WP MaddisonPE MidfordSP Otto2007Estimating a binary character's effect on speciation and extinction.Syst Biol56701710