Mar Biol (2014) 161:45–59 DOI 10.1007/s00227-013-2312-7 ORIGINAL PAPER Deconstructing an assemblage of “turtle” barnacles: species assignments and fickle fidelity in Chelonibia John D. Zardus · David T. Lake · Michael G. Frick · Paul D. Rawson Received: 3 May 2013 / Accepted: 8 August 2013 / Published online: 24 August 2013 © Springer-Verlag Berlin Heidelberg 2013 Abstract Barnacles in the genus Chelonibia are com- is genetically distinct—leading to the conclusion that the mensal with a variety of motile marine animals including three former taxa are morphotypes of the same species and sea turtles, crustaceans, and sirenians. We conducted a should be synonymized under C. testudinaria. Phyloge- worldwide molecular phylogenetic survey of Chelonibia netic analysis resulted in three geographic clades (Atlan- collected from nearly all known hosts to assess species tic, Indian Ocean/western Pacific, and eastern Pacific) with relationships, host-fidelity, and phylogeographic structure. haplotype parsimony networks revealing no shared hap- Using DNA sequences from a protein-coding mitochon- lotypes among geographic regions. Analysis of molecular drial gene (COI), a mitochondrial rRNA gene (12S), and variance detected significant differences among sequences one nuclear rRNA gene (28S), we found that of four spe- by region (p < 0.005); conversely, there were no signifi- cies, three (C. testudinaria, C. patula, and C. manati) are cant differences among sequences when grouped by host or genetically indistinguishable. In addition, we show each taxonomic designation. Average pairwise genetic distances utilizes a rare androdioecious mode of reproduction involv- were lower between the eastern Pacific and Atlantic clades ing complemental males. In contrast, the fourth species (C. (0.053 0.006) than between the eastern Pacific and ± caretta), which is hermaphroditic and specializes on turtles, Indian Ocean/western Pacific clades (0.073 0.008), sug- ± gesting Atlantic and eastern Pacific populations were con- nected more recently, perhaps until the rise of the Isthmus Communicated by C. Riginos. of Panama. Host use by Chelonibia morphotypes is dis- Electronic supplementary material The online version of this cussed along with speculation on possible ancestral hosts article (doi:10.1007/s00227-013-2312-7) contains supplementary and support for a “turtle-first” hypothesis. material, which is available to authorized users. J. D. Zardus (*) · D. T. Lake Department of Biology, The Citadel, 171 Moultrie Street, Introduction Charleston, SC 29407, USA e-mail: [email protected] Commensal associations between species, wherein one part- ner in the relationship benefits and the other is unaffected, Present Address: D. T. Lake are placed somewhere between mutualism and parasitism on University of South Carolina School of Medicine, the symbiosis spectrum (Leung and Poulin 2008; Frick and 6439 Garners Ferry Rd, Columbia, Pfaller 2013). Flexible host partnering occurs in facultative SC 29208, USA commensalisms, whereas obligate commensalisms involve M. G. Frick species-specific associations. Regardless of the degree of Archie Carr Center for Sea Turtle Research, Department of specialization in a particular host/symbiont relationship, it is Biology, University of Florida, Gainesville, FL 32611, USA increasingly recognized that symbiotic systems are impor- tant sources of evolutionary novelty (Sapp 1994; Vermeij P. D. Rawson School of Marine Sciences, University of Maine, 1994; Zook 2004) and commensalism may well be a com- 5751 Murray Hall, Orono, ME 04469, USA mon tipping point in the path to coevolution. 1 3 46 Mar Biol (2014) 161:45–59 Barnacles that are obligate associates of sea turtles, the and of low aspect when compared to C. patula (Fig. 1c), so-called turtle barnacles, present a compelling case for which typically has wedges with smooth edges at the examining the dynamics of commensalism. Balanomorph sutures, and is comparatively higher (taller) and thin- or acorn barnacles of the superfamily Coronuloidea include shelled. Chelonibia caretta (Fig. 1d) presents a robust approximately two dozen living species in several genera shell similar to, but not as thick as, C. testudinaria yet of and families that are best known for living attached to tur- higher aspect and with a uniform exterior surface lacking tles and whales (Pilsbry 1916; Newman and Ross 1976; wedges at the sutures. Monroe and Limpus 1979; Frick and Zardus 2010; Frick Pilsbry (1916) mentions phenotypic plasticity in Che- et al. 2010a, b; Ross and Frick 2011; Hayashi 2013). Per- lonibia, particularly in referring to C. manati-like forms haps the most frequently encountered coronuloids are removed from sea turtles, stating “there are certain bar- members of the genus Chelonibia Leach 1817, of which nacles in this series before me which, while possibly ref- five living species have been described, each documented erable to C. testudinaria as varieties, have some of the as commensal with a particular suite of hosts—ranging characters of the West African species [C. manti]…I am from crustaceans and chelicerates to sea turtles and sireni- giving names to these forms in order to call attention to ans (see Table 1). These barnacles are considered obligate their characters which might otherwise be overlooked by commensals with narrow host specialization. Some affili- those having opportunity of seeing large numbers of turtle ate with several host species, but no single barnacle species barnacles. Their status as races cannot yet be considered is reported to occur on all the documented hosts; though established.” And he refers to some fossil forms of Che- Chelonibia testudinaria (Linnaeus 1758) is reported from lonibia as “of the testudinaria type” that “afford no infor- all species of sea turtles. The enigmatic Chelonibia ramosa mation on the phylogeny of the genus.” He also remarks Korschelt 1933 is known only from a written description of that these differences seem to correspond to host selection a single individual found on an unspecified sea turtle (Kor- where C. testudinaria is “admirably adapted to the rough schelt 1933) and is, therefore, excluded from our study. conditions of existence on the backs of sea turtles, the walls The effects of chelonibiid barnacles on their hosts appear being enormously thickened and the stature low,” whereas generally benign, given the paucity of records reporting the relatively fragile and lighter C. patula is specialized negative effects and lack of correlation with turtle health for living on motile marine–estuarine animals, particularly (Stamper et al. 2005), but their presence in high numbers crabs. Moreover, Henry’s (1943) description of Chelonibia or in unusual locations (e.g., attaching in wounds and over- patula dentata (host unspecified) amalgamates characters growing eyes) certainly has adverse consequences for the of C. patula and C. testudinaria providing another indica- host (Zardus et al. 2007). Unlike rhizocephalan barnacles tor of phenotypic plasticity in the group. In examining large which are parasites in the strict sense of drawing nutrition numbers of barnacles, we also have observed Chelonibia from their hosts (Høeg 1995), turtle barnacles merely use specimens that mingle the characters of these “species” in a their basibont for substratum and transport. Benefits to the variety of ways. barnacles include increased dispersal, access to consist- Past accounts indicate that C. manati is the most specific ent feeding currents, and perhaps most importantly escape member of the genus—selecting manatees and occasionally from predators (Foster 1987). loggerhead sea turtles as hosts (but see Seigel 1983)—fol- Host specificity is the distinguishing feature differen- lowed by Chelonibia caretta (Spengler 1790) which tiating obligate from facultative commensalisms, and the attaches to only three of the seven extant species of sea various species of Chelonibia have historically been con- turtles. Less selective is C. testudinaria, occurring on all sidered obligate host commensals and described accord- sea turtles but also documented from saltwater crocodiles ing to their variation in morphology (Fig. 1). However, (Monroe and Garrett 1979), American alligators (Nifong they do display morphological variation within species, and Frick 2011), terrapins (Seigel 1983), laboratory glass- and when shell morphologies are not entirely consistent ware (Zardus and Hadfield2004 ), and slate and plastic or differ from established taxonomic characters, Che- settlement panels (Zardus, unpublished data). Chelonibia lonibia species are often identified by host type (Frick and patula (Ranzani 1818), historically defined as a crab bar- Ross 2001). Usually, C. manati Gruvel 1903 (Fig. 1a) can nacle, is perhaps least selective and has been recorded from be distinguished by its pleated shell plates, which pro- terrapins, (Ross and Jackson 1972), sea snakes (Badru- duce finger-like extensions from the basal edge that aid deen 2000), several different crustaceans (Ortiz et al. 2004; in grasping the flexible skin of sirenians. A common diag- Cheang et al. 2013), and inanimate substrata (Relini 1980; nostic trait of C. testudinaria is a stellate pattern on the Frazier and Margaritoulis 1990). A recent phylogenetic shell formed by open wedges at the sutures between shell study using many individuals of C. patula and C. testudi- plates, sculpted along their margins with indentations naria collected from the vicinity of the South China Sea or “teeth” (Fig. 1b). The shell