Molecular Phylogenetics and Evolution 67 (2013) 9–14

Contents lists available at SciVerse ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier .com/ locate /ympev

Short Communication

Phylogenetic position and evolutionary history of the turtle and whale barnacles(Cirripedia: Balanomorpha: Coronuloidea)

Ryota Hayashi a,b, Benny K.K. Chan c, Noa Simon-Blecher d, Hiromi Watanabe b, Tamar Guy-Haim d,Takahiro Yonezawa e, Yaniv Levy f, Takuho Shuto a, Yair Achituv d,⇑a Atmosphere and Ocean Research Institute, University of Tokyo, Kashiwa, Chiba 277-8564, Japanb Marine Biology and Ecology Research Program, Extremobiosphere Research Center, Japan Agency for Marine-Earth Science and Technology,Natsushimacho, Yokosuka 237-0061, Japanc Biodiversity Research Center, Academia Sinica, Taipei 115, Taiwand The Mina and Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat Gan 52900, Israele School of Life Sciences, Fudan University, 220, Handan Rd., 200433 Shanghai, Chinaf The Sea Turtle Rescue Center, Nature & Parks Authority. Mevoot Yam, P.O.B. 1174, Mikhmoret 40297, Israel

a r t i c l e i n f o

Article history:Received 31 May 2012Revised 16 December 2012Accepted 27 December 2012Available online 8 January 2013

Keywords:Turtle barnaclesCoronuloideaTetraclitoideaPhylogenyMolecular markersTime divergence

1055-7903/$ - see front matter � 2013 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.ympev.2012.12.018

⇑ Corresponding author. Fax: +972 3 7384058.E-mail address: yair.achituv@biu.ac.il (Y. Achituv).

a b s t r a c t

Barnacles of the superfamily Coronuloidea are obligate epibionts of various marine mammals, marinereptiles and large crustaceans. We used five molecular markers: 12S rDNA, 16S rDNA, 18S rDNA, 28SrDNA and Histone 3 to infer phylogenetic relationships among sixteen coronuloids, representing mostof the recent genera of barnacles of this superfamily. Our analyses confirm the monophyly of Coronuloi-dea and that this superfamily and Tetraclitoidea are sister groups. The six-plated Austrobalanus clusterswith these two superfamilies. Based on BEAST and ML trees, Austrobalanus is basal and sister to theCoronuloidea, but the NJ tree places Austrobalanus within the Tetraclitoidae, and in the MP tree it is sisterto both Coronuloidea and Tetraclitoidae. Hence the position of Austrobalanus remains unresolved. Withinthe Coronuloidea we identified four clades. Chelonibia occupies a basal position within the Coronuloideawhich is in agreement with previous studies. The grouping of the other clades does not conform to pre-vious studies. Divergence time analyses show that some of the time estimates are congruent with the fos-sil record while some others are older, suggesting the possibility of gaps in the fossil record.

� 2013 Elsevier Inc. All rights reserved.

1. Introduction

Barnacles of the superfamily Coronuloidea are obligate epi-bionts of various marine mammals, marine reptiles and crusta-ceans, attaching to the host‘s surface or embedded in their skin.Currently, this superfamily is divided into three families, Cheloni-biidae, Platylepadidae and Coronulidae (Newman, 1996). Super-family Coronuloidea is characterized by reduction of opercular,Xenobalanus is an extreme example in which opecular valves arelacking and the shell is reduced to a star shape attachment struc-ture embedded in the skin of its host.

The different opinions on the systematic position of the Coronu-loidea are presented in Supplementary material 1. Darwin (1854)classified the turtle and whale barnacles in the subfamily Balani-nae. Pilsbry (1916) (Supplementary material 1: 1A) applied familyand subfamily names to the Balanomorpha, within the Balanidae,he introduced two subfamilies, Chelonibiinae and Coronulinae, toinclude the turtle and whale barnacles. He assigned the genus

ll rights reserved.

Chelonibia to the Chelonibiinae and the other turtle barnacles andwhale barnacles to the Coronulinae. He proposed that there is nodirect relationship between these subfamilies and that the reducedopercular valves, presented in both subfamilies, are a convergentcharacter. Within the Coronulinae he proposed a ‘‘scheme’’ ofdivergence with two collateral ‘‘series’’. Newman and Ross (1976)upgraded Pilsbry’s two ‘‘series’’ to subfamilies (Coronulinae andPlatylepadinae) and included these subfamilies into the family Cor-onulidae together with Chelonibiinae and the extinct Emersoniinae(Supplementary material l; 1B). They also proposed phylogeneticrelationships within the Coronulidae. The basal position of Cheloni-bia is supported by its plesiomorphic characteristics. These includeretention of a separation between the rostrum and the rostro-lat-eral parietes, which in all other balanids form a compound ros-trum, possessing a membranous basis, lacking a basidorsal pointon the penis and having a weak bullate labrum. In addition, cirrusIII is of an intermediate form between cirrus II and IV. According totheir taxonomy the Coronulidae was included in superfamilyCoronuloidea along with the Tetraclitidae and the Bathylasmati-dae. The systematics of Newman and Ross (1976) was subse-quently modified by Newman (1996), elevating Newman and

10 R. Hayashi et al. / Molecular Phylogenetics and Evolution 67 (2013) 9–14

Ross’ (1976) subfamilies to the family level and organizing the tet-raclitids and bathylasmatids into the superfamily Tetraclitoidea.

Monroe (1981) used a ‘‘balloon’’ diagram to represent the sys-tematics of the turtle and whale barnacle (Supplementary materiall: 1C). His systematics was based on shell morphology and hostantiquity. In his systematics, the ancestral form is Chelonibia patula.He suggested that the Coronulidae contain four subfamilies: theChelonibiinae with one genus, Chelonibia; the Emersoniinae withone extinct genus, Emersonius; the Coronulinae with five generaand the Xenobalaninae with four genera. These relations are basedon morphological characters and on the ecological fact that they allare adapted to reside on live substrata. Ross and Frick (2011) fur-ther proposed four subfamilies in coronuloid barnacles, based onmorphological shell characteristics and revised the family-groupnames of the Coronuloidea. They proposed three families and 11subfamilies. All the subfamilies except Stomatolepadinae (Stoma-tolepas and Stephanolepas) and Coronulinae (Coronula, Cetopirus,Cetolepas) are monogeneric. Recently Hayashi (2012; 2013), em-braced Newman’s (1996) systematics however, these authors didnot suggest phylogenetic interpretation in their taxonomy.

These different taxonomies obscure the evolution of thecoronuloids. Our aim is to resolve the contradictions between thedifferent phylogenic opinions. We used two mitochondrial rRNAgenes, two nuclear rRNA genes and one protein-coding nuclearDNA gene as molecular markers. In our analyses we include repre-sentatives of the superfamilies Pachylasmatoidea, Chthamaloidea,Balanoidea and Tetraclitoidea following the systematic arrange-ment of Newman (1996).

2. Materials and methods

2.1. Taxon sampling

In this study samples were mainly collected from Japan and Is-rael. The material from Japan was obtained by catch or strandings.In Israel, barnacles were collected from turtles kept in the Sea Tur-tle Rescue Center, Mikhmoret, Israel. Additional material was ob-tained from the Caretta Research Project, Savvannah, Georgia,and from the collection of the Western Australia Museum, Perth.Chelonibia patula was removed from the carapace of a blue crabCallinectes sapidus from a fishmonger in Jaffa, Israel. Xenobalanusglobicipitis was removed from a false killer whale Pseudorca crassi-dens, in South Africa. Coronula diadema was removed from astranded whale. Cryptolepas rhachinecti was collected from astranded California Gray Whale (Cascadia Research group, Olym-pia, Washington). All necessary permits for collection and curationwere obtained. All samples were fixed and preserved in 95% etha-nol at �20 �C.

2.2. Outgroup selection

Verruca stromia was selected as an outgroup, based on the basalposition of Verrucomorpha within the Sessilia (Pérez-Losada et al.,2004, 2008). For divergence time estimation we used calibrationnodes 9–13 of Pérez-Losada et al. (2008) and their fossils ages.An additional calibration node was used in this study based onthe fossil record of Pachylasma veteranum from the Paleocene,65.5–55.8 Ma (Buckeridge, 1983).

2.3. DNA extraction, amplification, and sequencing

In our analysis sixteen taxa of Coronuloidea and five genera rep-resented in the fossil record were used to calibrate the moleculartree included. Total genomic DNA was extracted from barnaclemuscles and mantle tissue using the High Pure PCR template kit

(Roche, Germany). We avoided, when possible, using the cirriand soma to reduce contamination by food, and to enable a mor-phological examination if needed. DNA amplification and sequenc-ing were performed as described in Pérez-Losada et al. (2004).

We used the mitochondrial genes 12S rDNA, 16S rDNA, and thenuclear genes 18S rDNA, 28S rDNA and Histone 3 to infer relation-ships among taxa. Amplification products were sequenced for bothstrands at Macrogen Inc., Seoul (Korea). Sequences were manuallyinspected and edited using the BioEdit program (http://www.mbio.ncsu.edu/BioEdit) Sequences have been deposited inGenBank; accession numbers are given in Supplementary material2.

2.4. Phylogenetic analysis

The sequences were aligned using MAFFT v6 in the G-INS-imode (http://mafft.cbrc.jp/alignment/software/). Gblocks v0.91b(http://molevol.cmima.csic.es/castresana/Gblocks.html) was usedto exclude ambiguous positions (default settings except allowedgap = half). The sequences were concatenated to form a multi-genematrix including 28 taxa sequences, divided into five partitions,one for each gene. The jModelTest v0.1.1 program (http://dar-win.uvigo.es/software/software.html) was used in order to selectthe best fitting substitution model for each partition according tothe corrected Akaike information criterion (Supplementary mate-rial 3).

Maximum likelihood phylogeny of the concatenated data wasinferred with RAxML v7.2.8 (http://www.phylo.org/index.php/por-tal/) using a GTRCAT model of evolution with 50 rate categoriesand 1000 bootstrapping replicates. Bayesian analyses were con-ducted with MrBayes v3.1.2, (http://mrbayes.sourceforge.net/). Itwas not possible to implement the exact preferred model as se-lected by the AIC and therefore GTR + I + C substitution modelwas applied for all partitions. In both ML and Bayesian analyses,characters within combined sequence sets were partitioned bygene, allowing each partition to evolve at different rates. MrBayessearches were conducted with trees sampled every 1000 genera-tions. Analyses were run for three million generations. Conver-gence between runs and the choice of an appropriate burn-invalue were assessed by comparing the traces using Tracer v1.5(http://tree.bio.ed.ac.uk/software/tracer). A Neighbor-Joining (NJ)analysis was conducted using the Maximum Composite Likelihoodmethod and the rate variation among sites was modeled with agamma distribution. A Maximum Parsimony (MP) analysis wasconducted using the Close-Neighbor-Interchange algorithm. Theconsensus trees in the NJ and the MP analyses were inferred from1000 bootstrap replicates. Both analyses were conducted in MEGAv5.0 (www.megasoftware.net/).

2.5. Divergence time analyses

Including multiple calibration nodes has been shown to im-prove estimates of divergence times and rate estimates (Yang,2004; Porter et al., 2005; Pérez-Losada et al., 2008). Divergencetimes were therefore estimated using multiple genes and multiplefossil calibration points. A relaxed-clock MCMC approach using theuncorrelated log- normal model was implemented in BEAST v1.6.1(http://beast.bio.ed.ac.uk/), using 4 � 107 generations, and sam-pling every 1000th generation. The sequence data were dividedinto five partitions. Models of sequence evolution for each nucleo-tide sequence partition were determined using the correctedAkaike information criterion in jModelTest (Supplementary mate-rial 3). The Yule model was chosen as the speciation prior for allthree data sets.

Log files were analyzed using Tracer v1.5, to assess convergenceand to confirm combined effective sample sizes for all parameters

Table 2Divergence times (and 95% CI) for Coronuloidea and relatives as estimated using theBayesian evolutionary analysis method based on five genes and five fossilcalibrations.

Clade Taxon Divergence time (95% CI) (MYA)

Stem node Crown node

I C. manati + C. patula 9.009 (4.64–14.76)

3.514 (0.85–6.7)

II Austrobalanus imperator 65.225 (52.5–77.97)

–

III Chelonibia + Stephanolepas 52.973 (40.95–66.19)

30.721 (18.93–44.64)

IV Stomatolepas spp. 43.964 (31.07–56.78)

27.207 (15.24–41.19)

V Coronuloidea 65.225 (52.5–77.97)

60.721 (48.41–74.76)

VI Coronuloidea + Tetraclitoidea 72.162 (61.19–83.45)

65.225 (52.5–77.97)

R. Hayashi et al. / Molecular Phylogenetics and Evolution 67 (2013) 9–14 11

larger than 200. This was done to ensure that the MCMC chain hadrun long enough to obtain valid estimates of the parameters(Drummond and Rambaut, 2007). All resulting trees were thencombined with LogCombiner v1.6.1, (http://beast.bio.ed.ac.uk/)with a burn-in of 25%. A maximum credibility tree was subse-quently produced using TreeAnnotator v1.6.1 (http://beast.bio.e-d.ac.uk/).

2.6. Fossil calibrations

We based our divergence time estimates (Table 2) on five fossilcalibrations from closely related taxa, formerly applied by Pérez-Losada et al. (2008). The dates ranged from the Cretaceous (EarlyCretaceous, 99.6–65.5 Ma) to the Neogene (Early to Middle Mio-cene, 11.6–23 Ma). Since the fossils we used were dated to anage range, we used the midpoint of each geologic range for diver-gence time estimation (Buckeridge, 1983). The fossil age representsthe minimal age of an organism. Hence, it is more appropriate topresent a node within a time interval rather than a fixed time (Nor-ell, 1992). Calibrations were plotted to the node prior to the basalnode of the clade of interest.

3. Results and discussion

3.1. Phylogenetics and taxonomy

Our data set contained partial sequences of five genes from 16species of coronuloids, 11 representatives of Balanomorpha andVerruca stroemia as the outgroup. The concatenated data set is of4705 bps; the 12S rDNA sequence was 357 bps, 16S rDNA was419 bps, 18S rDNA was 1841 bps, 28S rDNA segment was 1761bps and Histone 3 was 327 bps long.

Our results based on four methods of phylogenetic analysis, NJ,MP, ML and Bayesian analyses (Fig. 1 and Supplementary material4), support the monophyly of the Coronuloidea There are nostrongly supported branches in conflict among the different phylo-genetic trees. Our analyses confirmed that Coronuloidea and Tetr-aclitoidea are sister groups within the Balanomorpha, as presentedby Newman and Ross (1976, Fig. 5) and Newman (1996). Themonophyly of the six-plated Austrobalanus with these two taxa isalso confirmed. However, Austrobalanus clusters differently amongthe trees. In the BEAST, ML and MP analyses it appears in a basalposition within the Coronuloidea. The bootstrap support of thisgrouping is 56% in ML tree with 0.74 Bayesian support, and 63%and 80% in MP and NJ trees respectively. This concurs with New-man and Ross (1976) who placed the Austrobalaninae at the baseof the Tetraclitoidea. The NJ tree is consistent with morphology,however ML and MP bootstrap support is low, therefore the posi-tion of Austrobalanus remains unresolved.

Within the Coronuloidea we distinguish four main clades: Che-lonibia which clusters with Stephanolepas, the single lineage ofStomatolepas, two species of Platylepas that cluster with Cylindrole-pas darwiniana and Tubicinella cheloniae and the whale barnacles

Table 1Species and ages of fossils used as calibrations for divergence time estimations (takenfrom Buckeridge, 1983).

Species Geological age (mya) Node

Pachydiadema (Catophragmus)cretacea

U. Cretaceous (Senonian) (70.6–89.3)

C1

Pachylasma veteranum Paleocene (65.5–55.8) C2Austromegabalanus victoriensis Neogene–M.–L. Miocene (11.6–23) C3Tetraclitella sp. cf.

purpurascensNeogene–L. Miocene (Aquitanian)(20.4–23.0)

C4

Chamaesipho brunnea Neogene–L. Miocene (16–23) C5

Coronula, Cryptolepas, and Xenobalanus which group with Cylind-rolepas sinica.

While, the basal position of Chelonibia within the Coronuloideaagrees with previous hypotheses based on its plesiomorphic char-acter, the inner grouping and division within the Coronuloidea donot conform to any schemes presented by other studies of thisgroup (Supplementary material 1). The clustering of Chelonibiaand Stephanolepas in the same clade is incongruent with all previ-ous studies based on morphology. Our results do not support theassumption that C. patula, which exploits phylogenetically olderhosts, such as crustaceans and molluscs as well as inanimate sub-strata, are the most ‘‘primitive’’ coronuloids (Ross and Newman,1967). Based on this proposition, Monroe (1981) suggested thatC. patula represents the ancestral condition of the Coronuloideaand those living on cetaceans are more ‘‘advanced’’. However,within the clade containing all the Chelonibia species, it is C. carettathat occupies the basal position. Chelonibia patula, regarded as anancestral form of the Coronuloidea by previous researchers (Rossand Newman, 1967; Monroe, 1981), was located at an end branchin the phylograms. This species clusters with C. manati with highsupport. These species are adapted to brackish environments livingon crabs, manatees and brackish water turtles (Ross and Jackson,1972; Seigel, 1983).

Newman’s (1996) Platylepadidae (Supplementary material 1:1B) is polyphyletic. The three species of Stomatolepas represent amonophyletic group of three OTUs with robust separation fromtheir other sister clades and they also have a clear separation with-in the Coronuloidea. Monroe’s (1981) Coronulinae (Coronula, Cry-ptolepas, Platylepas and Cylindrolepas; Supplementary material 1:1C) is both polyphyletic and paraphyletic since it does not includeXenobalanus. His Xenobalaninae which also includes Stephanolepas,Stomatolepas and Tubicinella appears to be polyphyletic. Also, thegenus Cylindrolepas was found to be polyphyletic. Cylindrolepasdarwiniana is grouped with Platylepas with strong bootstrap andBayesian support while C. sinica is clustered with the whale barna-cles. Species of Platylepas are found on sea turtles, sea snakes, dug-ongs and garfish while Cylindrolepas spp. are only found on seaturtles (Hayashi, 2009). Monroe (1981) regarded the genus Cylind-rolepas as a synonym of Platylepas because Cylindrolepas spp. looklike P. decorata and is embeded in host skin and have lateral ribson the parietes. Hayashi (2012) showed that the labrum of C. dar-winiana is multidentate like that of P. decorata and these speciesare also similar in their parietal wall. However, the labrum of C.sinica is not multidentate and has only few teeth on it. Multiplealigned sequences of several specimens of C. darwiniana, C. sinica,Platylepas and Tubicinella cheloniae (GenBank AB723950-AB723960; Supplementary material 5) supports the morphologicalobservations and clustering of these taxa. Monroe’s (1981) sugges-

Fig. 1. A. Molecular phylogenetic analysis of coronuloid taxa using the maximum likelihood estimation. The tree includes 28 species (16 coronuloids) and is based on fivepartitions, one for each of the following genes: 12S, 16S, 18S, 28S and H3 (total 4705 bp). At each node, the number before the slash indicates the percentage of ML bootstrapsupport (1000 replicates) from RAxML analysis applying the CTRCAT model. The number following the slash at each node indicates the Bayesian posterior probability, usingMrBayes, expressed as a decimal fraction for nodes that received at least 50% support in at least one analysis. Taxonomic units are linked with doted line. Hosts are indicated.B. Estimated Bayesian divergence time chronogram generated by BEAST v1.6.1 (Drummond and Rambaut, 2007), using the Maximum Clade Credibility consensus BMCMCtree from multi-locus molecular data. Clade posterior probabilities are shown for each node. Fossil calibration nodes are indicated by circled numbers C1–C5, correspondingwith Table 1. The outgroup, Verruca stroemia, is highlighted in gray. The major geologic periods are mapped onto the phylogeny using the following symbols: C, Cretaceous;PA, Paleocene; E, Eocene; O, Oligocene; M, Miocene; PL, Pliocene. Roman characters I–VI indicate clades for which divergence time was estimated (Table 2).

12 R. Hayashi et al. / Molecular Phylogenetics and Evolution 67 (2013) 9–14

R. Hayashi et al. / Molecular Phylogenetics and Evolution 67 (2013) 9–14 13

tion that Cylindrolepas is a junior synonym of Platylepas was notsupported and we have shown that both genera are distinct evolu-tionary significant units. In addition, our results did not supportthe division of the subfamily Stomatolepadinae and Cylindrolep-adinae, by Ross and Frick (2011).

The contradiction between the results based on molecular dataand the classic taxonomy based on morphology may stem from thefact that some morphological characteristics, which appear in dif-ferent clades within the Coronuloidea, are homoplasious charac-ters. Tubular or cylindrical shells are found in more than oneclade, as well as a cup or bowl shape shell. It appears that the epi-biotic mode of life lead to the evolution of these homoplasiouscharacteristics.

3.2. Divergence time analyses

Fig. 1B represents the estimated divergence time chronogramusing the Maximum Clade Credibility consensus BMCMC treebased on five genes, and five fossil calibrations. The divergencetimes and 95% highest posterior density (HPD) intervals are givenin Table 2. Multiple independent Bayesian runs produced higheffective sample sizes and convergence statistics in Tracer thatindicated all analyses had converged.

The age of sea turtle and whale fouling by balanomorphs is stillpoorly documented. Currently many taxa are missing from the fos-sil record (Hayashi, 2013) and therefore are marked Recent (New-man et al., 1969; Buckeridge, 1983). We estimated the age of thesebarnacle species and compared the inferred results with the pale-ontological record.

Our analyses suggest that Tetraclitidae originated in the earlyPaleocene, however, the recorded age is younger, dating to theMiocene (Newman et al., 1969). Austrobalanus also diverged fromCoronuloidea in the Paleocene, prior to the earliest fossil record ofthis taxon in the late Eocene (Buckeridge, 2000). Classically,Austrobalanus is regarded as presenting the most plesiomorphiccharacteristics in a clade that contains both the Tetraclitidaeand Coronuloidea (Fig. 1B, group VIII), while we found that it isrelated to the common ancestor of the Coronuloidea (Fig. 1B,group VII).

The oldest known coronuloid barnacle is the extinct Emersoniuscybosyrinx from the Late Eocene (Ross and Newman, 1967). AnEmersonius-like barnacle is interpreted as the ancestor of epizooticcoronuloid barnacles living on turtles or on other comparableorganisms (e.g. xiphosurans, large decapod crustaceans). Further-more, the oldest fossil record of a Chelonibia species is from the

Fig. 2. (A) Attachment traces of Platylepas hexastylos on a skull of loggerhead sea turtle, Csea turtle, C. caretta. (C) Enlargement of B. (D) Fossil sea turtle Euclastes melii from Mimodified from Misuri, 1910). Scale bars equal 50 mm.

Oligocene (Zullo, 1982). The estimated ages of diversification ofCoronuloidea and Chelonibia are Eocene and Oligocene respec-tively, and our results are congruent with these fossil records.

The age of Chelonibia based on morphological and moleculardata (Pérez-Losada et al., 2004) indicates that the separation ofthe Chelonibia clade from other balanomorphs was during the Cre-taceous. Our inferred age, based only on molecular data of sessil-ians barnacles, date Chelonibia to the early Oligocene, youngerthan the Cretaceous but older than the Miocene. The disagreementbetween the dates revealed by us and by Pérez-Losada et al. (2004)may result from the selection of reference nodes and the use of dif-ferent marker datasets, as we included only sessilans in our analy-sis. Recently, Harzhauser et al. (2011) described the extinctchelonibiid barnacle Protochelonibia submersa, a sister-group ofChelonibia, belonging to the Early Miocene. This finding is consis-tent with our results. Formerly, Chelonibia patula, which attachesto crustaceans, was regarded as one of the ancient forms of coronu-loids (Ross and Newman, 1967; Monroe, 1981). Although the fossilrecords of C. patula go back to the Late Miocene (Withers, 1929;Ross, 1963a), our analysis shows that it separates from its sisterspecies in the Pleistocene. The node of Clade I represents the diver-gence time from a common ancestor of C. patula and C. manati andthe establishment of an association between barnacles andSirenians.

Fossil records of Stomatolepas, Cylindrolepas, Tubicinella, Xeno-balanus and Stephanolepas were not found in earlier fossils andwere therefore dated as Recent (Buckeridge, 1983) or Pleistocene(Foster and Buckeridge, 1987) offshoots of the Coronulidae. How-ever, based on our data, we found them to be older. The estimatedage of Stomatolepas (Fig. 1B, group IV), and the two groups: group V(Tubicinella, Cylindrolepas darwiniana and Platylepas, Fig. 1B), andgroup VI (Cylindrolepas sinica, Xenobalanus, Cryptolepas and Coron-ula, Fig. 1B) is Oligocene. That of Xenobalanus (Fig. 1B) and of Steph-anolepas is Miocene. The recorded age of Coronula within theMiocene is confirmed. The inferred age of Cryptolepas is the Mio-cene while the recorded age is in the Late Pleistocene (Zullo, 1961).

According to our divergence time analyses, the species of Platy-lepas were derived in the Early to Middle Miocene. However, thefossil record of the Platylepas, P. wilsoni, is from the Pleistocene(Ross, 1963b). The rarity of fossil records of this genus might bedue to the fragility of its parietal walls. Platylepas hexastylos, havingcharacteristic wall plates with midribs, leaves a characteristic im-print of its attachment on turtle bone (Fig. 2A–C), trace of Platyle-pas like barnacle can be detected on a figure of fossil sea turtle,Euclastes melii from the Miocene found in Pietra Leccese, Southern

aretta caretta. (B) Attachment trace of P. hexastylos on a carapace bone of loggerheadocene Pietra Leccese in Southern Italy with barnacle trace on the carapace (photo

14 R. Hayashi et al. / Molecular Phylogenetics and Evolution 67 (2013) 9–14

Italy (Misuri, 1910) (Fig. 2D), this is congruent with our estimateddivergence time of Platylepas.

Epibiotic organisms are potential targets for understanding co-evolution and our findings are a step in the direction of revealingthe origin of such a relationship. The evolutionary history of seaturtles, whales and barnacles, including the use of trace fossils,should be combined to reveal co-evolutionary patterns.

Acknowledgements

The Japanese team thanks the many people who generouslyhelped in various ways during the course of this study. These in-cludes the fishermen of the Kanna and Toya fishery port in Oki-nawa, the Hahajima fishery port in Tokyo and the Hidejima setnet fishermen in the Cooperative Association of Miyako in Iwatefor providing live sea turtles. The team would also like to thankthe Sea Turtle Association of Japan, the Nippon-Koei Co., Ltd. Oki-nawa office, the Entomological laboratory and Fujukan of Univer-sity of Ryukyus, the Ever Lasting Nature of Asia, the Club NOAHHahajima, Ogasawara Marine Center, Chura-Umi Aquarium of Oki-nawa, the Seikai National Fishery Research Institute of IshigakijimaIs., the National Museum of Nature and Science, Tokyo, the Insti-tute of Cetacean Research of Japan, the Historical Geology and Pal-aeontology laboratory of Chiba University and the numerousvolunteer staff of Yakushima Umigame-kan who aided RH in bothfield work and observations. We additionally thank Kunio Komesu,Muneyuki Kayo, Kouichi Hirate, Kazuki Tsuji, Hidetoshi Ota, Muts-umi Kaneko, Katsufumi Sato, Shingo Kimura, Junichi Okuyama, Ta-dasu K. Yamada, Yuko Tajima, Akiko Yatabe, Kenichiro Fujita,Masayuki Ishii, Masahiro Aizawa, Takeharu Bando, Yukihisa Omu-ta, Ren Hirayama, and Urara Kuratani for aiding this research andproviding study equipments and materials. This study was sup-ported by the program ‘Bio-Logging Science of the University of To-kyo (UTBLS)’.

This study was supported by grant 574/10 of the Israel ScienceFoundation (ISF). The Israeli team would like to thank Dr. M. Frickfrom the Caretta Research Project, Savvannah, Georgia, Prof. KeithCrandal and Dr. Marcos Pérez-Losada from BYU, Provo and CIBIOPortugal and to Dr. Jessica Huggins from Cascadia Research group,Olympia, Washington for help in providing samples of turtle andwhale barnacles.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.ympev.2012.12.018.

References

Buckeridge, J.S., 1983. Fossil barnacles (Cirripedia: Thoracica) of New Zealand andAustralia. N. Z. Geol. Surv Paleontol. Bull. 50, 1–151, +plates.

Buckeridge, J.S., 2000. A new species of Austrobalanus (Cirripedia, Thoracica) fromEocene erratics, Mount Discovery, McMurdo Sound, East Antarctica.Paleobiology and Paleoenvironments of Eocene Rocks, McMurdo Sound. EastAntarctica. AGU Antarctic Res. Ser. 76, 329–333.

Darwin, C., 1854. A Monograph on the Subclass Cirripedia, with Figures of all theSpecies. The Balanidae, the Verrucidae, etc.. Ray Society, London.

Drummond, A., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis bysampling trees. BMC Evol. Biol. 7, 214.

Foster, B.A., Buckeridge, J.S., 1987. Barnacle palaeontology. In: Southward, A.J. (Ed.),Barnacle Biology, Crustacean Issues, vol. 5. A.A. Balkema. Rotterdam, TheNetherlands, pp. 43–63.

Harzhauser, M., Newman, W.A., Grunert, P., 2011. A new Early Miocene barnaclelineage and the roots of sea-turtle fouling Chelonibiidae (Cirripedia,Balanomorpha). J. Syst. Paleontol. 9, 473–480.

Hayashi, R., 2009. New host records of the turtle barnacle, Cylindrolepas sinica: acase study of sea turtles’ behaviour and their epibionts. Marine BiodiversityRecords 2, e165.

Hayashi, R., 2012. Atlas of the barnacles on marine vertebrates in Japanese watersincluding taxonomic review of superfamily Coronuloidea (Cirripedia:Thoracica). J. Mar. Biol. Assoc. UK 92, 107–127.

Hayashi, R., 2013. A checklist of turtle and whale barnacles (Cirripedia: Thoracica:Coronuloidea). J. Mar. Biol. Assoc. UK 93, 143–182.

Misuri, A., 1910. Sopra un nuovo chelonio del calcare miocenico di Lecce (Euclastesmelii Misuri). Palaeontographia Ital. 16, 119–136.

Monroe, R., 1981. Studies on the Coronulidae (Cirripedia): shell morphology,growth, and function, and their bearing on subfamily classification. Mem.Queensland Mus. 20, 237–251.

Newman, W.A., 1996. Cirripedia; Suborders Thoracica and Acrothoracica. In: Forest,J. (Ed.), Traité de Zoologie, vol. 7, no. 2. Masson, Paris, Crustacés, pp. 453–540.

Newman, W.A., Ross, A., 1976. Revision of the balanomorph barnacles, including acatalog of the species. Mem. San Diego Soc. Nat. Hist. 9, 1–108.

Newman, W.A., Zullo, V.A., Withers, T.H., 1969. Cirripedia. Treat. Invertebr.Paleontol. Part R, Arthropoda 4, 206–295.

Norell, M.A., 1992. Taxic origin and temporal diversity: the effect of phylogeny. In:Novacek, M.J., Wheeler, Q.D. (Eds.), Extinction and Phylogeny. ColumbiaUniversity Press, New York, pp. 89–118.

Pérez-Losada, M., Høeg, J.T., Crandall, K.A., 2004. Unraveling the evolutionaryradiation of the Thoracican barnacles using molecular and morphologicalevidence. A comparison of several divergence time estimation approaches. Syst.Biol. 53, 244–264.

Pérez-Losada, M., Harp, M., Hoeg, J.T., Achituv, Y., Jones, D., Watanabe, H., Crandall,K.A., 2008. The tempo and mode of barnacle evolution. Mol. Phyl. Evol. 46, 328–346.

Pilsbry, H.A., 1916. The sessile barnacles (Cirripedia) contained in the collections ofthe US National Museum: including a monograph of the American species. Bull.US Natl. Mus. 93, 1–366.

Porter, M.L., Pérez-Losada, M., Crandall, K.A., 2005. Model-based multi-locusestimation of decapod phylogeny and divergence times. Mol. Phyl. Evol. 37,355–369.

Ross, A., 1963a. Chelonibia in the Neogene of Florida. Quart. J. Florida Acad. Sci. 26,221–233.

Ross, A., 1963b. A new Pleistocene Platylepas from Florida. Quart. J. Florida Acad. Sci.26, 150–158.

Ross, A., Frick, M.G., 2011. Nomenclatural emendations of the family-group namesCylindrolepadinae, Stomatolepadinae, Chelolepadinae, Cryptolepadinae, andTubicinellinae of Ross & Frick, 2007—including current definitions of family-groups within the Coronuloidea (Cirripedia: Balanomorpha). Zootaxa 3106, 60–66.

Ross, A., Jackson Jr., C.G., 1972. Barnacle fouling of the ornate diamondback terrapin,Malaclemys terrapin macrospilota. Crustaceana 22, 203–205.

Ross, A., Newman, W.A., 1967. Eocene Balanidae of Florida, including a new genusand species with a unique plan of ‘‘Turtle-Barnacle’’ organization. Am. Mus.Novitates 2288, 2–21.

Seigel, R.A., 1983. Occurrence and effects of barnacle infestations on diamondbackterrapins (Malaclemys terrapin). Am. Midl. Nat. 109, 34–39.

Withers, T.H., 1929. The cirripede Chelonibia in the Miocene of Gironde, France andVienna, Austria. Annal. Mag Nat. Hist. ser.10. 4, 566–569.

Yang, Z., 2004. A heuristic rate smoothing procedure for maximum likelihoodestimation of species divergence times. Acta Zool. Sinica 50, 645–656.

Zullo, V.A., 1961. A new whale barnacle from the Late Pleistocene deposits at SanQuintin Bay, Baja California. Veliger 4, 13–14.

Zullo, V.A., 1982. A new species of the turtle barnacle Chelonibia Leach, 1817,(Cirripedia, Thoracica) from the Oligocene Mint Spring and Byram Formationsof Mississippi. Mississippi Geol. 2, 1–6.