Reversible shifts between interstitial and epibenthic habitats in evolutionary history: Molecular phylogeny of the marine flatworm family Boniniidae (Platyhelminthes: Polycladida: Cotylea) with descriptions of two new species

Tiny animals in various metazoan phyla inhabit the interstices between sand and/or gravel grains, and adaptive traits in their body plan, such as simplification and size reduction, have attracted research attention. Several possible explanations of how such animals colonized interstitial habitats have been proposed, but their adaptation to this environment has generally been regarded as irreversible. However, the actual evolutionary transitions are not well understood in almost all taxa. In the present study, we show reversible evolutionary shifts from interstitial to epibenthic habitats in the lineage of the polyclad flatworm genus Boninia. In addition, we establish two new species of this genus found from different microhabitats on a single beach in Okinawa Island, Japan: (i) the interstitial species Boninia uru sp. nov. from gravelly sediments and (ii) the epibenthic species Boninia yambarensis sp. nov. from rock undersurfaces. Our observations suggest that rigid microhabitat segregation exists between these two species. Molecular phylogenetic analyses based on the partial 18S and 28S rDNA sequences of the new Boninia species and four other congeners, for which molecular sequences were available in public databases [Boninia antillara (epibenthic), Boninia divae (epibenthic), Boninia neotethydis (interstitial), and an unidentified Boninia sp. (habitat indeterminate)], revealed that the two interstitial species (B. neotethydis and B. uru sp. nov.) were not monophyletic among the three epibenthic species. According to ancestral state reconstruction analysis, the last common ancestor of the analyzed Boninia species inhabited interstitial realms, and a shift to the epibenthic environment occurred at least once. Such an “interstitial to noninterstitial” evolutionary route seems to be rare among Animalia; to date, it has been reported only in acochlidian slugs in the clade Hedylopsacea. Our phylogenetic tree also showed that the sympatric B. uru sp. nov. and B. yambarensis sp. nov. were not in a sister relationship, indicating that they colonized the same beach independently rather than descended in situ from a common ancestor that migrated and settled at the beach.

Introduction Animals inhabiting the space between sand and/or gravel grains have attracted the attention of biologists since the 1930s [cf. 1], primarily due to their miniaturized body size [e.g., [2][3][4][5][6] and ecological importance [e.g., 7], although the existence of such tiny animals was recognized by zoologists in the 19th century [e.g., 8,9]. An assemblage of such animals is known as interstitial fauna, a term first introduced in 1935 to refer to small copepods, nematodes, rotifers, and protozoans found at sandy beaches [10]. In addition, the term mesopsammon (literally meaning "between sand") has been used since the 1940s, initially and chiefly in German literature [11]. Other terms, such as meiobenthos and meiofauna, are also often used interchangeably with interstitial fauna. Technically, however, meiobenthos (or meiofauna) are organisms that can pass through a 1-mm mesh but are retained by a 45-μm mesh. Therefore, interstitial or mesopsammic animals are not necessarily meiobenthic in size.
To date, interstitial animals have been documented in at least 23 of the~34 currently recognized metazoan phyla [12], and different evolutionary scenarios have been proposed to explain the existence of such animals [e.g., 4]. Recent phylogenetic studies have shown that some annelid interstitial taxa had independently derived from larger epibenthic ancestors either by progenesis or stepwise miniaturization depending on the taxa [6,13]. Additional research on these evolutionary processes has shed light on other taxa, including Enteropneusta (Hemichordata) [14], Acochlidiacea (Heterobranchia: Gastropoda) [15], Rhodopemorpha (Heterobranchia: Gastropoda) [16], and Ostracoda [17]; however, the vast majority of relevant animal groups have yet to be studied in this context, including the phylum Platyhelminthes and one of its constituent subtaxa, the order Polycladida.
According to recent phylogenomic analyses [18,19], Polycladida is reciprocally monophyletic to another order, Prorhynchida, which together form a clade that is sister to the remaining platyhelminths, excluding Catenulida and Macrostomorpha. All the known catenulids and macrostomorphans are microscopic, and some species live in interstitial habitats [20], whereas some almost exclusively mesopsammic flatworm groups, such as Proseriata and Rhabdocoela, are more deeply nested in the phylum [18,19].
Little attention has hitherto been paid to interstitial polyclads, probably due to their rarity. Indeed, polyclads were not mentioned in Swedmark's 1964 seminal work on marine interstitial fauna [21]. Polyclads are mostly free-living marine flatworms that are categorized into two suborders: Acotylea (with 29 families [22]) and Cotylea (with 12 families [cf. 23]). Most polyclads have a relatively large body size (�5 mm) and dwell on the surface of the marine bottom [24], whereas only a tiny fraction inhabit interstitial environments [25]; of the~800-1,000 species of polyclads described worldwide to date, 12 (5 in 4 acotylean families and 7 in 3 cotylean families) are mesopsammic in the adult stage [25][26][27][28][29][30]. On the other hand, some surface dwellers can also be found from interstitial environments when they are juveniles or subadults (authors' personal observation). In this paper, we restrict the notion of interstitial polyclads to refer to those that inhabit sand/gravel sediments even after reaching sexual maturity. Likewise, epibenthic and noninterstitial species refer to the ones inhabiting environments other than the interstitial one, such as undersurfaces of rocks, when they are fully mature, while juveniles or subadults may dwell in sand or gravel.
The cotylean polyclad family Boniniidae Bock, 1923 [31] is interesting in terms of the evolutionary shift between noninterstitial and interstitial habitats because it harbors members that live in both environments. However, the phylogenetic inter-relationships within this family are yet to be resolved because of insufficient taxon sampling [23]. Currently, seven named species of boniniids are classified into three genera: Boninia Bock, 1923 [31] (5 species), Paraboninia Prudhoe, 1944 [32] (1 species), and Traunfelsia Laidlaw, 1906 [33] (1 species). To date, only Boninia neotethydis Curini-Galletti & Campus, 2007 [30] from the Mediterranean and Red Sea has been described as a permanent interstitial representative based on adult specimens [30]. Boniniids are morphologically characterized by having (i) a narrow and elongate body with a pair of pointed tentacles located at the anterolateral margins, (ii) a male copulatory apparatus that includes an unarmed penis papilla and one or several prostatoid organ(s) with stylets, (iii) a female copulatory apparatus with a Lang's vesicle, and (iv) a ventral adhesive organ located at the posterior end of the body [24].
During our polyclad faunal survey in Japan, we found two undescribed species of Boninia on a single beach, with one species collected from an interstitial environment and the other from rock undersurfaces. From these findings, we hypothesized some evolutionary scenarios pertaining to (i) shifts between interstitial and noninterstitial microhabitats and (ii) settlement of the two species at the same beach. Of the conceivable hypothetical scenarios, one suggests the two interstitial species of Boninia (B. neotethydis in the Mediterranean/Red Sea and the undescribed form in Japan) being exclusively monophyletic. This "interstitial monophyly hypothesis" would be supported if adaptation from a noninterstitial to interstitial lifestyle was evolutionarily irreversible and uncommon. In another hypothesis, the last common ancestor of our two new species, which could be either interstitial or noninterstitial, settled at the beach, and one of the two species subsequently changed microhabitat. This "in situ speciation hypothesis" would be favored if such an event was considered rare that the settlement of two closely related species (i.e., in the same genus) at a single beach happened twice independently.
Overall, the aims of this study were to (i) provide formal taxonomic descriptions of the two new Boninia species and (ii) test the abovementioned hypotheses using ancestral state reconstruction analysis based on a molecular phylogenetic tree of Boniniidae.

Ethics statement
No permissions were required for collecting materials in this study. Our sampling locality was not privately owned but open to the public. We did not involve endangered or protected species.

Collection of specimens and morphological observations
Specimens were collected at Nagahama Beach, Okinawa Island, Japan. Gravelly sediment samples near the edge of the water were agitated in tap water to extract animals. The supernatant was filtered using an about 1-mm meshed dip net, and the residue was subsequently transferred into seawater. In total, six polyclads were extracted from the sediment samples. Other six polyclads crawling on undersurfaces of rocks were also collected at the sandy beach in the intertidal area. All flatworm specimens were anesthetized in a MgCl 2 solution prepared with tap water to match the seawater salinity using an IS/Mill-E refractometer (AS ONE, Japan), after which they were photographed using a Nikon D5600 digital camera with external strobe lightning provided by a pair of Morris Hikaru Komachi Di flash units. Each polyclad specimen was fixed and preserved using one of the four protocols shown in Table 1.
The fixation protocols (i-iv) are: (i) a part of the body was removed and preserved in 99.5% ethanol for DNA extraction and the rest of the body was fixed in Bouin's solution for 24 h and subsequently preserved in 70% ethanol; (ii) the whole body was fixed in Bouin's solution for 24 h and subsequently preserved in 70% ethanol; (iii) the whole body was preserved in 99.5% ethanol for DNA extraction; and (iv) the whole body was mounted on a glass slide, squeezed under a cover slip, and preserved in 10% formaldehyde solution in seawater.
For histological examination, tissues were prestained with acid fuchsin, dehydrated in an ethanol series, cleared in xylene, embedded in paraffin wax, and sectioned serially at a thickness of 4 μm using a microtome. Sections were stained using either hematoxylin and eosin (HE) or Mallory's trichrome (MT) methods, mounted on glass slides, and then embedded in Entellan New (Merck, Germany) under cover slips. They were then observed and photographed using a Nikon D5300/5600 digital camera under an Olympus BX51 compound microscope.
Type specimens have been deposited in the Invertebrate Collection of the Hokkaido University Museum (ICHUM). All graphical treatments were completed using Adobe Photoshop CC. Illustrations were prepared using Adobe Illustrator CC.

DNA extraction, PCR, and sequencing
Total DNA was extracted using a DNeasy Blood & Tissue Kit (Qiagen, Germany) after specimens were kept overnight at 55˚C in 180 μl of ATL buffer (Qiagen, Germany) with 20 μl of proteinase K (>700 U/ml; Kanto Chemical, Japan). As a reference for DNA barcoding, a partial sequence (709 bp) of the cytochrome c oxidase subunit I (COI) gene was determined. For phylogenetic inference, fragments of the 18S rDNA (18S; 1,758 bp) and 28S rDNA (28S; 1,014-1,015 bp) were sequenced. Amplification of the three gene markers was performed using polymerase chain reaction (PCR) via a 2720 Thermal Cycler (Applied Biosystems, USA); 10-μl reaction volumes were used, each of which contained 1 μl of total DNA template, 1 μl of 10 × ExTaq buffer (Takara Bio, Japan), 2 mM of each dNTP, 1 μM of each primer, and 0.25 U of Takara Ex Taq DNA polymerase (5 U/μl; Takara Bio, Japan) in deionized water. The forward and reverse primer pairs listed in Table 2 were used. The PCR amplification conditions were as follows: 94˚C for 5 min; 35 cycles of 94˚C for 30 s, 50˚C (18S and COI) or 52.5˚C (28S) for 30 s, and 72˚C for 2 min (18S), 1.5 min (28S), or 1 min (COI); and 72˚C for 7 min. PCR products were purified enzymatically using ExoSAP-IT reagent. All nucleotide sequences were determined using direct sequencing with a BigDye Terminator Kit ver. 3.1 and a 3730 Genetic Analyzer (Life Technologies, California, USA) with the primers listed in Table 2. Sequences were checked and edited using MEGA ver. 7.0 [38]. All edited sequences have been deposited in DDBJ/EMBL/GenBank with accession numbers LC699268-LC699282 (Table 1).

Phylogenetic analyses
For phylogenetic analyses, a concatenated dataset (2,685 bp) comprising partial 18S (1,739 bp) and 28S rDNA (946 bp) sequences was prepared. Additional 18S and 28S rDNA sequences of four species from Boniniidae, which were available in a public database, were downloaded from GenBank (Table 3). To assess the last common ancestral state of boniniids, its proposed sister groups Amyellidae Faubel, 1984 [37] and Theamatidae Marcus, 1949 [23, 26, 41] were also included in the analysis ( Table 2). The three cotylean species Cestoplana rubrocincta (Grube, 1840) [42], Pericelis flavomarginata Tsuyuki et al., 2020 [43], and Pericelis tectivorum Dittmann et al., 2019 [44] were used as outgroups (Table 2). Sequences were aligned using MAFFT ver. 7.427 [45], with the L-INS-i strategy selected using the "Auto" option. Ambiguous sites were trimmed using Clipkit ver. 1.0 via the "kpic" option [46]. The optimal substitution models, selected using PartitionFinder ver. 2.1.1 [47] according to the Akaike Information Criterion [48] with the greedy algorithm [49], were GTR+I+G for both the 18S and 28S partitions. Phylogenetic analysis was performed using the maximum likelihood (ML) method via RAxML ver. 8.2.10 [50]. Bayesian inference (BI) of the phylogeny was performed using MrBayes ver. 3.2.3 [51,52] with two independent runs of Metropolis-coupled Markov chain Monte Carlo (MCMC), each consisting of four chains of 2,000,000 generations. All parameters (statefreq, revmat, shape, and pinvar) were unlinked between each position; trees were sampled every 100 generations. The first 25% of the trees were discarded as burn-in before a 50% majority-rule consensus tree was constructed. Convergence was confirmed using an average standard deviation of split frequencies of 0.001989, potential scale reduction factors for all parameters of 1.000-1.002 and effective sample sizes for all parameters of >209. Nodal support within the ML tree was assessed using analysis of 1,000 bootstrap (BS) pseudoreplicates [53]. ML BS values �70% and posterior probability (PP) values �90% were considered to indicate clade support (here, combined nodal support is indicated as "PP/BS").

Ancestral state reconstruction related to microhabitat
The habitat of each ingroup species was determined from the original description (Table 3). The habitat information of the three unidentified species, Boninia sp., Chromyella sp., and Theama sp., was provided directly by the collector, Christopher Edward Laumer ( Table 3). The possible ancestral states were reconstructed using Bayesian Binary MCMC (BBM) analysis implemented in RASP 4.2 [56,57]. To take phylogenetic uncertainty into account, 10 trees randomly selected from the post burn-in trees generated by MrBayes ver 3.2.3 were used as input trees. BBM analysis was then run on a consensus Bayesian tree. The MCMC chain was run for 50,000 generations using 10 chains and sampled every 100 generations. A fixed (LC) model that did not allow null root distribution was used to conduct the analysis. The specimen is currently registered as B. antillara based on the taxon concept of Litvaitis et al. [23] in that B. divae should be synonymized with B. antillara, but it was originally identified as B. divae based on the morphology [23]. b In the GenBank database, the specimen was assigned to Boninia divae, but it should be "Boninia sp." because it was unidentifiable due to its juvenile state (cf. https:// mczbase.mcz.harvard.edu/guid/MCZ:IZ:132897). c Although the specimen was collected from an interstitial habitat, we treated the habitat of Boninia sp. as indeterminate in this paper because we cannot evaluate the habitat in adult state due to its juvenile state (see Introduction). https://doi.org/10.1371/journal.pone.0276847.t003

Nomenclatural acts
The electronic vision of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved, and the associated information viewed through any standard web browser by appending the LSID to the prefix "http://zoobank.org/." The LSID for this publication is: urn:lsid:zoobank.org:pub:B88724EB-7332-419E-A4C4-A1DFC05E121F. The electronic edition of this work was published in a journal with an ISSN and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS, ResearchGate, HUSCAP. Etymology. The specific name uru, an Okinawan dialect meaning "coarse sand," is derived from the habitat of the species.

Taxonomy
Diagnosis. Body narrow and elongated; one pair of pointed tentacles located at anterolateral margins; four cerebral eyespots and 21-29 marginal eyespots; 2-4 prostatoid organs arranged into single girdle; Lang's vesicle fully ciliated; subepidermal muscle fibers of adhesive area not well developed.
Tentacular eyespots absent. Pair of two cerebral eyespots (ca. 24 μm in diameter) located at each anterior side of brain; two eyespots in each pair lying close to each other (Fig 1C). Marginal eyespots (ca. 55 μm in diameter), 21-29 in number (27 in holotype), distributed sparsely in anterior quarter of body along margins on both sides (Fig 1A-1C). Diameter of marginal eyespots twice as large as that of cerebral eyespots (Fig 1C).
Adhesive organ located at posterior end of body on ventral side (Figs 1B, 1D and 3I). Subepidermal muscle fibers not well developed in adhesive area, surface of which are covered by thick glandular epithelium (Fig 3I).
Distribution. To date, known only from the type locality: Nagahama Beach, eastern coast of Okinawa Island, Japan.
Habitat. To date, confirmed only from gravelly habitats in intertidal coarse sediments. Remarks. Our specimens are assigned to Boninia because they conform to the generic diagnosis of Curini-Galletti & Campus [30], i.e., they have two or more prostatoid organs with stylets opening into the male atrium. Boninia uru sp. nov. can be easily distinguished from B. antillara, B. divae, and B. mirabilis by its single girdle of prostatoid organs [30,31] (Table 4). The other two congeners B. neotethydis and B. oaxaquensis have a single girdle of prostatoid organs, as in the new species; however, B. uru sp. nov. is distinguishable from these two species by its small number (2-4) of prostatoid organs (10-18 organs in B. neotethydis and 16-24 organs in B. oaxaquensis). Additionally, the arrangement of eyespots enable discrimination

PLOS ONE
Reversible shifts between interstitial and epibenthic habitats in evolutionary history of Boniniidae  [30] is conspecific with the new species described here. Boninia sp. was collected from Samboanga and originally identified as B. mirabilis by Bock [31]. Later, Curini-Galletti & Campus [30] re-examined Bock's [31] voucher specimens and recognized them as an undescribed species based on their internal morphology, including (i) the very small number (3-7) of prostatoid organs arranged into a single girdle and (ii) the completely unciliated Lang's vesicle. The new species is similar to the specimens from Samboanga in terms of the small number of prostatoid organs and the single girdle, but it differs by its entirely ciliated Lang's vesicle. Etymology. The new species is named after the region Yambaru, the northern part of Okinawa Island. The type locality, Nagahama Beach, is located in the southeastern Yambaru region.

Boninia yambarensis
Diagnosis. Body narrow and elongated; pair of pointed tentacles located at anterolateral margins; 3-4 pairs of cerebral eyespots and 19-42 marginal eyespots; 21-22 prostatoid organs arranged into single girdle; five uterine vesicles present in each oviduct; Lang's vesicle fully ciliated; subepidermal muscle fibers of adhesive area not well developed.
Adhesive organ located at posterior end of body on ventral side (Figs 4C and 6F). Subepidermal muscle fibers not well developed in adhesive area. Distribution. The species is known from the type locality, Nagahama Beach, eastern coast of Okinawa Island, Japan.
Habitat. To date, confirmed only from under rocks in the intertidal region. The thin body width (1-2 mm) of this species suggests that it may also be able to inhabit intergravel spaces. However, B. yambarensis sp. nov. seems to have a preference for epibenthic habitats over interstitial habitats because (i) it has yet to be collected from interstitial environments and (ii) more than 10 individuals of the species were found under rock surfaces independently in our two surveys.
Remarks. The materials examined belong to Boninia because they conform to the generic diagnosis, i.e., they have two or more prostatoid organs with stylets opening into the inner area of the male atrium. Boninia yambarensis sp. nov. can be separated from B. antillara, B. divae, and B. mirabilis by its single girdle of prostatoid organs [30,31]. Boninia yambarensis sp. nov. resembles B. neotethydis, B. oaxaquensis, and B. uru sp. nov. in having a single girdle of prostatoid organs (Table 4)

Molecular phylogeny
The resulting BI and ML trees were identical in terms of topology; all six species of Boninia exclusively formed a clade (0.99 PP; 95% BS) (Fig 7).  (Fig 7).

Ancestral habitats
The ancestral states of habitats reconstructed via BBM analysis are shown in Fig 8. The last common ancestor (LCA) of all analyzed species, including the outgroups (node 11), was estimated to be epibenthic with a probability of 98.7%. The LCAs of Boninioidea sensu Dittmann et al. [41] (node 9) and Boniniidae (node 5) appeared to be interstitial, although the estimated probabilities were relatively low (56.1% and 58.2%, respectively). In contrast, the LCA of Boninia sp., B. antillara, and B. divae (node 2) and that of Boninia sp. and B. antillara (node 1) were epibenthic with high probabilities of 97.5% and 99.8%, respectively. Also, the ancestral states of nodes 3 and 4 were likely to be epibenthic, which was the most favored state (53.4% and 60.3%, respectively).

Discussion
Our phylogenetic results suggest the possibility that an unexpected evolutionary scenario occurred in the Boninia lineage. Boninia uru sp. nov. was sister to a clade composed of the remaining five congeners, in contrast to both our stated hypotheses (see the Introduction), in which this new species would have been sister to the interstitial B. neotethydis ("interstitial monophyly hypothesis") or to the sympatric B. yambarensis ("in situ speciation hypothesis"). Thus, in this section and based on our results, we discuss the most plausible evolutionary  hypotheses pertaining to (i) the shift between interstitial and noninterstitial microhabitats and (ii) the settlement of the two species at the same beach.

Reversible evolutionary shifts from interstitial to epibenthic realms in the Boninia lineage
The results of our ancestral state reconstruction analysis show that early boniniids likely lived in interstitial microhabitats, with some descendants subsequently having evolved to inhabit epibenthic environments, whereas others either remained in (Fig 8). The LCA of all analyzed Boninia species (node 5) was estimated as interstitial, although this estimation is not supported with high probability (58.2%) (Fig 8). In contrast, the LCA of Boninia sp., B. antillara, and B. divae (node 2) and that of the former two species (node 1) were estimated to be epibenthic with high support (97.5% and 99.8%, respectively). These results suggest an evolutionary scenario in which the LCA of all analyzed Boninia species inhabited an interstitial environment, and where the LCA of Boninia sp., B. antillara, and B. divae subsequently changed to an epibenthic lifestyle.
A prerequisite for this interpretation is that microhabitat preference of adults is species-specific and alternative, i.e., boniniids in the same species do not occur simultaneously in both interstitial and noninterstitial environments at random in their mature state. We consider this assumption to be realistic and applicable based on our observations. In our three independent field surveys, we collected six individuals of B. uru sp. nov. only by washing gravel sediments near the highwater limit where rocks were absent. In contrast, on the same beach, we observed >10 individuals of B. yambarensis sp. nov. crawling on undersurfaces of rocks in the lower intertidal zone. These observations indicate a narrow habitat range at least for each new species described herein (see also habitat for B. yambarensis above). Such a habitat preference would be expected for the other analyzed species B. antillara, B. divae, and B. neotethydis by extrapolating the empirical evidence observed in our two new species, although the actual microhabitat for each of the other congeners should be confirmed in additional investigations in the future.
The evolutionary shift from interstitial to noninterstitial habitats is likely uncommon among Animalia. Indeed, irreversible one-way transition from the noninterstitial realm to the interstitial realm seems to be the norm among metazoan taxa investigated to date; such interstitial taxa are exclusively monophyletic, e.g., Dinophilidae, Diurodrilidae, Polygordidae, Protodrilidae, Psammodrilidae [60] (Annelida), Ototyphlonemertidae [61,62] (Nemertea), and Rhodopemorpha [63] (Mollusca), with the notable exception of acochlidean slugs in the clade Hedylopsacea [15]. Moreover, even among acochlidian slugs, evolutionary transitions from interstitial to noninterstitial habitats are limited to species living in specialized habitats, such as those exposed to nonmarine salinities (brackish, limnic, and amphibious species) [15] and living in the deep sea [64], whereas almost all other species of acochlidian slugs live in shallow waters. Our study suggests a habitat shift from the interstitial to noninterstitial marine realm in the evolutionary history of flatworms based on molecular phylogenetic evidence with statistical support.
It remains unclear what makes such unique evolutionary transitions from interstitial to noninterstitial habitats possible in the Boninia lineage. The relatively high phenotypic plasticity in adult body size (about >2-10 times) among polyclads [cf. 65] might be related to the evolutionary pathway. As Westheide [4] stated, body size is one of the most important factors for microhabitat shifts between interstitial and noninterstitial realms. In acochlidians, "secondary gigantism" in body size (see [4,66]) may have contributed to the evolutionary shift from interstitial to epibenthic habitats; secondary gigantism is likely to be a consequence of adaptation to brackish water, freshwater, and terrestrial systems [15] or to limitations of food resources in the deep sea [64]. If interstitial boniniids show plasticity in body size, accidental "gigantism" could potentially have led to a lifestyle outside interstitial biotopes, similar to the known example in acochlidians.

Independent colonization of the same beach
Our tree topology suggests that B. uru sp. nov. and B. yambarensis sp. nov. settled at the same beach independently. In the resulting tree, B. yambarensis sp. nov. was more closely related to the Caribbean and Lessepsian species (B. antillara, B. divae, B. neotethydis, and Boninia sp. of Laumer & Giribet [54]) than to the sympatric species B. uru sp. nov. (Fig 7). Additionally, the two new species clearly differ morphologically in their reproductive organs, i.e., the number of prostatoid organs (2-4 in B. uru sp. nov.; 21-22 in B. yambarensis sp. nov.). Thus, there seems to be deep divergence between the two new species, and they may have encountered the collection site after they had been reproductively isolated.