Variations of the radula and shell microstructures in 33 species of Japanese chiton were investigated along with molecular phylogenetic trees. The molecular phylogenetic trees indicated that Chitonida was composed of four clades, of which two clades formed Acanthochitonina and corresponded to Mopalioidea and Cryptoplacoidea, respectively, and the other clades formed Chitonina. In the radula, the shapes of the central and centro-lateral teeth and the petaloid process varied greatly among species or genera and were useful for the identification of particular species or genera. The presence of accessory and petaloid processes and the cusp shape were relatively conserved and useful for recognizing particular genera or even suborders. In the valves, four to six shell layers were found at the section, but the ventral mesostracum was not observed in Acanthochitonina. The shell microstructures in the ventral sublayer of the tegmentum varied at suborder, but those in the other layers were almost constant. The megalaesthete chamber type varied at superfamily and was helpful to identify particular families or superfamilies. The characteristics of the shell layers and shell microstructures appear to be a synapomorphy shared by the members of Acanthochitonina. The classification within Chitonina needs to be reexamined because the variations of the cusp shape and megalaesthete chamber type were relatively large and did not correspond to the current classification. Callochiton formed a sister group with Chitonida and would be equally closely related to Chitonina and Acanthochitonina because of possessing a mosaic of characteristics from both.
INTRODUCTION
Chitons, one of the minor groups in Mollusca, are elongate-oval and flattened in shape and have eight shell valves like armor. These valves are surrounded by a girdle, and the foot is on the ventral side. They have no head, and the oral region lacks eyes; instead, sensory organs called aesthetes are embedded in the valves. Chitons inhabit only seawater, clinging to the surface of something solid such as rocks and dead shells, and most eat algae or minute organisms with the radula. Their fossil records may date back to the Late Cambrian (Smith, 1960; Runnegar et al., 1979; Puchalski et al., 2008; Pojeta et al., 2010), and an obvious chiton fossil would be Echinochiton from Ordovician (Pojeta et al., 2003; Pojeta and Dufoe, 2008).
Approximately 1200 species of chitons have been discovered from shallow to deep seas in the world (Eernisse and Reynolds, 1994; Slieker, 2000; Schwabe, 2005; MolluscaBase editors, 2021). Chitons have been mainly classified based on the characteristics of shell, girdle, and radula (Bergenhayn, 1955; Kaas and Van Belle, 1980, 1990, 1994). Furthermore, Sirenko (2006) indicated a new classification by adding characteristics such as gills, glands, egg hull projections, and spermatozoids. The phylogenetic relationships of chitons have been examined using these, and other characters, such as DNA sequences (Okusu et al., 2003; Sigwart et al., 2010, 2013; Irisarri et al., 2014, 2020), aesthete canals and complexes (Fernandez et al., 2007; Vendrasco et al., 2008), and shell microstructures (Peebles et al., 2017). These phylogenetic relationships predominantly corresponded to each other in the higher taxa, but sometimes not in the lower taxa.
The characteristics of radulae are useful tools for the classification of gastropods (e.g., Sasaki, 2010; Ponder and Lindberg, 2020), and those of shell microstructures reflect the phylogenetic relationships of higher taxa, especially in bivalves (e.g., Taylor et al., 1969, 1973; Carter, 1990). As for chitons, Dall (1879) described the radula dentition of North Pacific chitons with comparing species. Thiele (1909–1910) provided monographs of worldwide chitons and emphasized the characteristics of the radula more than other chiton researchers. Sirenko has been actively publishing more than 100 publications covering most chiton taxa, and the majority of them include scanning electron microscope (SEM) images of the radula (e.g., Sirenko and Zhang, 2019; Sirenko, 2020). For Japanese species, Taki Is and Taki Iw published precise illustrations of the radula without SEM (e.g., Taki Is and Taki Iw, 1929a; Taki Is, 1938). Saito (2004) analyzed the radula variations of many species in Cryptoplacoidea and reported that their characteristics were useful in identifying the species and genus and could be used to construct the phylogenetic relationships. Bergenhayn (1930) was the first to examine the shell microstructures of chitons, but the descriptions were not very detailed because SEM was not used. Recently, Peebles et al. (2017) described the mineralogy and shell microstructures of eight chiton species collected from New Zealand using SEM and discussed the phylogenetic relationships at family level. However, in both Saito (2004) and Peebles et al. (2017), the phylogenetic resolutions were not so clear. The characteristics of the radula alone were not enough to analyze the higher taxa, and the shell microstructures alone were not enough to analyze the lower taxa. Probably, in the phylogenetic analysis of chitons, combining the characteristics of the radula and shell microstructures is necessary.
More than 90 species of chitons have been discovered on or near the coasts of Japan. The classification of Japanese chitons was well arranged by Taki Is and Taki Iw (1929a, 1929b, 1929c, 1930, 1931a, 1931b), Taki Is (1938), and Saito (1994, 1995, 1998, 2000, 2001, 2006, 2011, 2017). Most Japanese chitons can be identified according to the taxonomic keys created by them.
Table 1.
Species names, sampling localities, sampling dates, and accession numbers. Newly or renewedly determined DNA sequences are in bold.
In the present study, Japanese chitons living in the shallow sea were investigated. They inhabit various environments, and the species diversity is relatively high. First, phylogenetic relationships were constructed using the DNA sequences of mitochondrial cytochrome c oxidase subunit I (COI) and 16S ribosomal RNA (16S), and nuclear 18S ribosomal RNA (18S) and 28S ribosomal RNA (28S). Radula and shell microstructures were observed by SEM. Based on these results, the relationships among the variations of the radula and shell microstructures and the molecular phylogenetic tree of Japanese chitons living in the shallow sea are discussed.
MATERIALS AND METHODS
Sampling
All investigated specimens, identified as 33 chiton species, were collected from the adjacent water of Japan. They were collected from rocks or boulders on the shore during low tide, except Leptchiton aequispinus (Bergenhayn, 1993) collected at a water depth of 210 m by dredging, and Lept. hakodatensis (Thiele, 1909), Ischnochiton paululus Is. Taki, 1938, Tonicella lineata (Wood, 1815) and Cryptochiton stelleri (Middendorff, 1847) collected at a water depth of 3–8 m by SCUBA diving. The dates and localities of the samplings were shown in Table 1 and Fig. 1. All collected specimens were preserved in 100% EtOH directly or after freezing at –20°C. The species identification was performed as described by Saito (2017), and their classification followed the MolluscaBase editors (2021).
Phylogenetic analysis
Total DNA was extracted from the foot or rectal muscle of each specimen using a DNeasy Blood & Tissue Kit (QIAGEN). COI, 16S, 18S, and 28S gene regions were amplified by polymerase chain reaction (PCR) using Premix Taq (Takara) and a thermal cycler (T100, Bio-Rad). As for the 18S and 28S gene regions, two fragments that were separated into the upper and lower regions were amplified. The primers used are listed in Table 2. The conditions for PCR amplification were as follows: denaturation at 94°C for 30 s; annealing at 48°C (COI), 55°C (16S and 18S), or 58°C (28S) for 30 s; and extension at 72°C for 60 s. These steps were repeated 25–35 times. The PCR products were purified using ExoSAP-IT (Affymetrix), and cycle-sequencing reactions were performed using a BigDye Terminator v3.1 Cycle Sequencing Kit (ABI). DNA sequences were analyzed using a genetic analyzer (3130 Genetic Analyzer, ABI) in both 5′ and 3′ directions. All sequences were registered in the DNA Data Bank of Japan (DDBJ) (Table 1).
The alignment of DNA sequences for each gene region was performed using MAFFT v7.475 (Katoh et al., 2005). Sites those contained gaps or were of questionable homology were trimmed using trimAl v1.4rev15 (Capella-Gutiérrez et al., 2009). A molecular phylogenetic tree was constructed from the concatenated sequence of the four gene regions by Maximum Likelihood (ML) and Bayesian methods. The ML method was performed using RAxML-NG v1.0.2 (Kozlov et al., 2019). A bootstrap test was performed 10,000 times. The model for the ML method was selected for the sequence of each gene region using ModelTest-NG v0.2.0 (Darriba et al., 2020). The Bayesian method was performed using MrBayes v3.1.2 (Ronquist and Huelsenbeck, 2003). This program was run for 50,000,000 generations sampling every 1000th generation. The model for the Bayesian method was selected for the sequence of each gene region using Kakusan4 (Tanabe, 2011). Convergence in the Bayesian method was assessed using Tracer v1.7.1 (Rambaut et al., 2018).
Fig. 1.
Sampling localities. Abashiri: 44°03′N, 144°16′E; Shari: 44°03′N, 144°57′E; Akkeshi: 43°01′N, 144°49′E; Hakodate: 41°45′N, 140°43′E; Asamushi: 40°54′N, 140°51′E; Zushi: 35°16′N, 139°34′E; Manazuru: 35°09′N, 139°09′E; Misaki: 35°10′N, 139°32′E; Shimoda: 34°40′N, 138°56′E; Isso: 30°27′N, 130°30′E; Onna: 26°29′N, 127°50′E.
Table 2.
Primers used in the present study. Superscript numbers show the combination in PCR amplification.
Radula
The radula was carefully extracted from the soft parts of each specimen, which were fixed in 100% EtOH or frozen at –20°C. Each radula was cleaned with a dissecting needle under a binocular microscope (Olympus, SZ40), washed with an ultrasonic washer (Velvo-Clear, VS-100III), and preserved in 100% EtOH. Afterward, the radula was coated with 50-nm-thick platinum by ion sputter (JEOL, JEC-3000FC) and observed using SEM (JEOL, JCM-5000) at 10 kV accelerating voltage. The terminology for the radula followed Saito (2004) and Schwabe (2010).
Shell microstructures and aesthete canals and complexes
The investigated specimen was soaked in 10% chlorine bleach solution (Kao, Kitchen Haiter) for 1–4 days to remove the soft parts. The remaining eight valves were washed with pure water and desiccated at 60°C for 24 h. Valves III–VI were fractured by hand or with a micro-chisel and hammer. The sections along the antero-posterior axis were coated with 50-nm-thick platinum by ion sputter (JEOL, JEC-3000FC) and observed using SEM (JEOL, JCM-5000) at 10 kV accelerating voltage. The shell layers were identified as described by Connors et al. (2012); however, the anterior and posterior myostracums were not distinguished. The shell microstructures were identified according to Carter (1990). The aesthete canals and complexes were identified on the basis of Baxter and Jones (1981, 1984) and Currie (1992).
Fig. 2.
Molecular phylogenetic trees. Clade A: some Chitonina, including the Chitoninae part of Chitonidae; Clade B: the Toniciinae plus Acanthopleurinae part of Chitonidae; Clade C: Mopalioidea; Clade D: Cryptoplacoidea. Filled circles on each node indicate bootstrap > 50% in the ML tree and posterior probability > 0.90 in the Bayesian tree, and open circles indicate not so.
RESULTS
Molecular phylogenetic tree
The sequence lengths for the COI, 16S, 18S, and 28S gene regions were 557, 505–516, 1692–1709, and 1279–1313 base pairs (bp), respectively. After alignment and trimming, the analyzed lengths were 557, 499, 1693, and 1283 bp, respectively. The model for the ML method was GTR + G4 for the 16S, 18S, and 28S, TIM3 + I + G4 for the first base of the COI, TVM + I + G4 for the second base, and TIM2 + I + G4 for the third base. The likelihood index was -ln 26478.320061. The model for the Bayesian method was HKY85 + G for the 16S and the third base of the COI, SYM + G + I for the 18S, and GTR + G + I for the 28S and the first and second bases.
The ML and Bayesian trees are shown in Fig. 2. The topologies of the two trees mostly corresponded to each other, and the bootstrap values and posterior probabilities on the nodes were significantly high. However, there were a few differences in the positions of Ischnochiton spp. and Cryptoc. stelleri. In both trees, Chitonida was monophyletic and contained four clades. Clade A was composed of some Chitonina, including the Chitoninae part of Chitonidae and actually including Ischnochiton, Rhyssoplax, Stenoplax, Tripoplax, Lepidozona, and Callistochiton. Clade B was composed of the Toniciinae plus Acanthopleurinae part of Chitonidae and including Acanthopleura, Liolophura, Lucilina, and Onithochiton. Clade C was composed of Mopalioidea and including Tonicella, Mopalia, Cryptochiton, Placiphorella, and Schizoplax. Clade D was composed of Cryptoplacoidea and including Acanthochitona, Notoplax, and Cryptoplax. Ischnochitonidae, Chitonidae, Mopaliidae, and Acanthochitonidae were polyphyletic, and additionally, Ischnochiton and Acanthopleura were not monophyletic. Callochiton formed a sister group with Chitonida.
Fig. 3.
SEM images of radula. A: Leptochiton aequispinus; B: Lept. hakodatensis; C: Callochiton foveolatus; D: Ischnochiton comptus; E: Rhyssoplax komaiana; F: Stenoplax alata; G: I. hakodadensis; H: Tripopla albrechti; I: Lepidozona coreanica; J: Callistochiton jacobaeus; K: Acanthopleura gemmata; L: Acanthop. loochooana; M: Lucilina amanda; N: Lu. lamellosa; O: Onithochiton hirasei; P: Tonicella lineata; Q: To. zotini; R: Mopalia retifera; S: Cryptochiton stelleri; T: Placiphorella stimpsoni; U: Schizoplax brandtii; V: Acanthochitona defilippii; W: Notoplax conica; X: Cryptoplax japonica. a: central tooth; b: centro-lateral (first) tooth; c: major (second) lateral tooth; d: petaloid process; e: accessory process. Scale bar unit indicates micrometer.
Radula
SEM images of the representative radulae of the investigated specimens are shown in Fig. 3. The characteristics of the radula were more or less different among the species; however, these differences were relatively moderate within the genus, except for Ischnochiton hakodadensis Carpenter, 1893. The central teeth had various shapes, but they were grouped into six types: bulbous, cup-shaped, deltoid, hooked-needle, rectangular, and spatulate. The centro-lateral (first) teeth also varied, but they were grouped into six types: columnar, envelope-like, fingerlike, fungiform, sphenoid, and winglike. In addition, the accessory process, which is a minute process on the centro-lateral tooth (Saito, 2004), was observed in members of clades A and B, and Lept. aequispinus. The major (second) lateral teeth were grouped into four types: unicuspid, bicuspid, tricuspid, and patellar. Although the type of the major lateral tooth largely varied in clade A, it was conserved in the other clades. Furthermore, the petaloid process, which is a process below the head of the major lateral tooth (Thiele, 1893; Bullock, 1988; Saito, 2004), was found in clades A and B and was grouped into three types: cornered, nodule, and securiform. The characteristics of the radula of each species are shown in Table 3.
Shell microstructures
In the investigated specimens belonging to Callochiton, clades A and B, and the outgroups, six shell layers (i.e., tegmentum, dorsal mesostracum, articulamentum, ventral mesostracum, myostracum, and hypostracum) were observed in the valves. As for clades C and D, five shell layers, excluding the dorsal mesostracum from the six layers, were observed. Except in Cryptoc. stelleri, the number of shell layers was constant within the clade. In the valve of Cryptoc. stelleri, only four layers (i.e., articulamentum, ventral mesostracum, myostracum, and hypostracum) were found (Fig. 4).
The tegmentum of Callochiton, clades A and B, and Placiphorella stimpsoni (Gould, 1859) was composed of two layers. The dorsal sublayer was homogeneous (Ho) structure, and the ventral sublayer was mainly composite prismatic (CP) structure but partly Ho structure (Fig. 5A–H). In clades C and D, and the outgroups, the tegmentum was only one layer, the Ho structure (Fig. 5I–L). The dorsal and ventral mesostracums were constructed with crossed-lamellar (CL) structure (Fig. 5M–P). The articulamentum was composed of two layers in all investigated species. The dorsal sublayer was Ho structure, and the ventral sublayer was mainly CP structure but partly Ho structure (Fig. 5Q, R). In Cryptoc. stelleri, the dorsal sublayer often contained irregular simple prismatic (ISP) structure (Fig. 5S). The myostracum in all investigated species was observed between the ventral mesostracum and the hypostracum and/or beneath the hypostracum. This layer was mainly constructed with Ho structure (Fig. 5T, U) but sometimes contained ISP structure (Fig. 5V). The hypostracum was constructed with CL structure (Fig. 5W, X). The apophyses, the inner shell layers extending beyond the tegmentum, were composed of three layers: articulamentum, ventral mesostracum, and myostracum (Fig. 4).
Table 3.
Characteristics of radula.
Aesthete canals and complexes
Multiple-branch aesthete canals, which are a type of aesthete canals and complexes and widely distributed in all valves (Baxter and Jones, 1981, 1984; Currie, 1992), were observed in the tegmentum of all investigated species, except Cryptoc. stelleri. The shapes of the megalaesthete chambers in the aesthete complexes were observed and grouped into four types (Fig. 6). Type A had a cylindrical shape shortened more or less and was observed in members of clades A, B, and D. Type B had a sprawling shape into which micraesthete canals often merged before entering and was observed only in clade C. Type C had a wider shape in the middle and was observed in Callochiton and members of clades A and B. Type D had a bud-like shape with a stalk-like aesthete canal and was observed only in the outgroups. Table 4 lists the characteristics of the shell layers, shell microstructures, and megalaesthete chamber type.
The above results were listed and shown in relation to the phylogenetic tree in Fig. 7.
Fig. 4.
Schematic of valve section. A: section along antero-posterior axis for species of Callochiton and clades A and B; B: clades C and D; C: Cryptochiton stelleri. D, E, F: close view of the box in the figure. In Placiphorella stimpsoni, the tegmentum was composed of two layers, homogeneous and composite prismatic structures.
DISCUSSION
Molecular phylogeny
The ML and Bayesian trees in the present study were highly robust, though the phylogenetic positions of Ischnochiton spp. and Cryptoc. stelleri were unclear. The topologies of these trees were predominantly similar to those in Irisarri et al. (2014, 2020), which reported that Chitonoidea (Chitonina) formed a sister group with a clade comprising Mopalioidea plus Cryptoplacoidea (Acanthochitonina), but not to those in Okusu et al. (2003). Additionally, the phylogenetic position of Callochiton in the present study corresponded to those in Sigwart et al. (2010), Irisarri et al. (2014, 2020), and Moles et al. (2021), but not to that in Sigwart et al. (2013), which described that Callochiton formed a sister group with a clade comprising Mopalioidea plus Cryptoplacoidea. The phylogenetic position of Cryptoc. stelleri in the present study made Acanthochitonidae and Mopaliidae polyphyletic groups. This position corresponded to those reported by Kelly and Eernisse (2008) and Irisarri et al. (2014), and the latter study considered Cryptoc. stelleri to be a member of Mopaliidae. Therefore, the present study also grouped Cryptoc. stelleri into Mopaliidae (Figs. 2, 7). Ischnochiton hakodadensis, the only species possessing an accessory process in the genus, made Ischnochitonidae a polyphyletic group, as reported by Owada (2016, 2018). Furthermore, the present study clarified that Rhyssoplax, which was classified as Chitonidae by Sirenko (2006), formed a sister group with Ischnochiton. However, the phylogenetic position of Rhyssoplax did not correspond to that reported by Irisarri et al. (2020). The DNA sequence of Rhyssoplax used by Irisarri et al. (2020) was determined by Riesgo et al. (2012). If their identification was correct, it would be suggested that Rhyssoplax is a polyphyletic group. The present study also clarified that Lepidozona coreanica (Reeve, 1847), a common species along the coast of Japan, formed a sister group with Callistochiton jacobaeus (Gould, 1859) (Figs. 2, 7). These results indicate the need to reexamine the classification within Chitonina.
Fig. 5.
SEM images of shell microstructures. Dorsal and ventral sublayers of tegumentum in (A, B) Callochiton foveolatus, (C, D) Ischnochiton comptus, (E, F) Liolophura japonica, (G, H) Placiphorella stimpsoni. Tegmentum in (I) Leptochiton aequispinus, (J) Tonicella zotini, (K) Schizoplax brandtii, (L) Acanthochitona defilippii. Dorsal mesostracum in (M) Callo foveolatus, (N) Stenoplax alata. Ventral mesostracum in (O) Sc. Brandtii, (P) Cryptoplax japonica. Dorsal and ventral sublayers of articulamentum in Lucilina lamellose (Q, R). Dorsal sublayer of articulamentum in Cryptochiton stelleri (S). Myostracum between the ventral mesostracum and hypostracum in Lepidozona coreanica (T). Myostracum beneath the hypostracum in (U) Tonicella lineata, (V) Rhyssoplax komaiana. Hypostracum in (W) I. boninensis, (X) Acanthoc. achates. Ho: homogeneous structure; CP: composite prismatic structure; CL: crossed-lamellar structure; ISP: irregular simple prismatic structure. Scale bar indicates 10 µm. The upper direction of SEM images is dorsal side, and the lower direction is ventral side.
Fig. 6.
Megalaesthete chamber type in multiple-branch aesthete canals. (A) type A in Acanthopleura gemmate, (B) type B in Placiphorella stimpsoni, (C) type C in Ischnochiton comptus, (D) type D in Leptochiton hakodatensis. ac: aesthete canal, mc: megalaesthete chamber. White arrowhead indicates the aperture of megalaesthete canal, and black arrowhead indicates that of micraesthete canal. Scale bar indicates 100 µm in (A), 50 µm in (B, C, D).
Table 4.
Characteristics of shell layers, shell microstructures, and megalaesthete chamber type.
Radula
Saito (2004) suggested that the characteristics of the radula in chitons were useful not only for species identification but also for the assessment of the phylogenetic relationships at higher levels, emphasizing examples of Cryptoplacoidea in his comparisons. The present study indicated that the characteristics of the radula could identify the species (Table 3, Fig. 7), and support Saito (2004). The shapes of the central and centro-lateral teeth varied among species and genera and could be used for identification at least Japanese chiton species at those levels. Likewise, the shape of the petaloid process is informative at genus level, and considering all these tooth shapes helps to confirm identification. In contrast, the presence of accessory and petaloid processes and the major lateral cusp shape were found to be relatively conservative, but the cusp shape can be somewhat useful for distinguishing genera.
Fig. 7.
Relationships among the molecular phylogenetic trees, the genera or species, the character state of radula, shell microstructures and megalaesthete chamber type, and the classification in the higher taxon. p.: patellar, Ho: homogeneous structure, CP: composite prismatic structure.
Thiele (1893) described the petaloid process as a diagnostic feature for groups corresponding to Chitonoidea in the present study, but until now it had not been tested with molecular phylogenetic analysis. Based on molecular phylogenetic analysis, the present study indicated that the petaloid process only occurs in Chitonina. The accessory process is likely convergent in Chitonina and Lept. aequispinus (Fig. 7) because it was not found in other Lepidopleurida, Callochitonida, or Acanthochitonina within Chitonida.
Shell microstructures
Bergenhayn (1930) reported that chiton valves consisted of four shell layers (i.e., tegmentum, articulamentum, hypostracum, and mesostracum), excluding the periostracum. However, Haas (1972, 1976, 1981) did not distinguish between articulamentum and hypostracum and described chiton valves as being constructed from three shell layers (i.e., tegmentum, hypostracum, and myostracum). Poulicek and Kreusch (1986) supported Bergenhayn (1930) but did not recognize mesostracum, and Carter (1990) supported Haas (1972, 1976, 1981). Connors et al. (2012) defined seven shell layers: tegmentum, dorsal mesostracum, articulamentum, ventral mesostracum, anterior and posterior myostracums, and hypostracum; Tonicella marmorea (O. Fabricius, 1780) was an exception in lacking the dorsal mesostracum. The present study adopted the definition of Connors et al. (2012) but did not distinguish between anterior and posterior myostracums because they did not appear to vary in the assembled information. Peebles et al. (2017) reported that the valves of chitons were composed of four to seven shell layers according to Connors et al. (2012) and suggested that the properties of the shell layers did not clearly reflect the phylogenetic relationships because of the large species variations. The present study indicated that the valves were constructed of four to six layers and that the tegmentum and articulamentum had sublayers. Additionally, the present study has extended the observation of Connors et al. (2012) that the dorsal mesostracum layer is absent in To. marmorea; it is also absent in the seven other Acanthochitonina genera, and other species of Tonicella examined in the present study (Fig. 7). The congruent distribution relative to the molecular result implies that a shared absence of the dorsal mesostracum layer is a synapomorphy of Acanthochitonina (Table 4, Fig. 7).
Differences in the shell microstructures were observed only in the ventral sublayer of the tegmentum. In Callochiton, Chitonina, and P. stimpsoni, the sublayer was mainly constructed with CP structure, and with Ho structure in the others. It is likely that a clade comprising Callochiton and Chitonida first obtained the Ho structure in the sublayer because the two species of Leptochiton, which are outgroups, had Ho structure. Presumably, the CP structure was independently acquired in Callochiton, Chitonina, and P. stimpsoni. In Cryptoc. stelleri, the structure of the valve differed greatly from the others because the tegmentum was not observed. This may be because the valves are always embedded in the soft part. Peebles et al. (2017) distinguished the CL structure in the mesostracum or hypostracum into at least two types, based on the size of the fibers that made the CL structure. However, the present study did not distinguish such types of CL structures because the difference between CP and Ho structures is much larger than that between the types of CL structures. Furthermore, individual variation in the size of the fibers was often observed.
Aesthete canals and complexes
Baxter and Jones (1981, 1984) and Currie (1992) observed aesthete canals and complexes in detail, and Fernandez et al. (2007) and Vendrasco et al. (2008) additionally constructed the phylogenetic relationships using their characteristics. Their phylogenetic conclusions are consistent with the present study. In the present study, the megalaesthete chamber type, which was applied as one of the characters for the phylogenetic analysis in Fernandez et al. (2007) and Vendrasco et al. (2008), was available for identification of particular families or superfamilies. However, in Chitonina, the variation in the type was relatively high compared with the others. Although the megalaesthete chamber type may be related to the life type, both Acanthopleura living on the surface of a rock in the intertidal zone and having a lens on the surface of the valve, and Lepidozona living on the back of a rock in the subtidal zone and lacking a lens, possessed the type A.
Conclusions
In the radula, the shapes of the central and centro-lateral teeth and the petaloid process were useful for the identification of particular species or genera. The presence of accessory and petaloid processes and the cusp shape were available for recognizing particular genera or even suborders. In the valves, the number of shell layers and the shell microstructure type in the ventral sublayer of tegmentum appear to be a synapomorphy shared by the members of suborder Acanthochitonina, and the megalaesthete chamber type was helpful for identifying particular families or superfamilies. These characteristics of the valve could also be useful for verifying the fossil records of chitons. The molecular phylogenetic tree implied that the above characteristics were significant and simultaneously indicated that the classification within Chitonina would need to be reexamined because the variations of the cusp shape and megalaesthete chamber type did not correspond to the current classification. Callochiton, which is contained in Callochitonida in the current classification, would be equally closely related to any member of Acanthochitonina and any member of Chitonina because of possessing a mosaic of characteristics from both. Buckland-Nics and Hodgson (2000) investigated the characteristics of the gametes in Callochiton castaneus (W. Wood, 1815) treated as Callochiton dentatus (Spengler, 1797) in MolluscaBase editors (2021), and reported that Callo. castaneus was basal to Chitonida and closely related to both Chitonina and Acanthochitonina. This result is consistent with the present study.
The radula and shell microstructures have been broadly conserved in chitons living in the shallow sea. Especially, shell microstructure characters are among the relatively few morphological features that can be observed in fossil chitons. They were found to be phylogenetically informative at both geologically ancient and recent levels.
ACKNOWLEDGMENTS
I deeply thank Professor Akiya Hino (Kanagawa University) and Professor Hiroshi Hosoya (Kanagawa University) for letting me use the laboratory for DNA sequencing. I am grateful to Associate Professor Michiko Nemoto (Okayama University), Dr. Masaya Saito (Kanagawa University), and Mr. Shoichi Hamano (Hokkaido University) for helping to collect the samples. I am indebted to Professor Douglas J. Eernisse (California State University Fullerton) for fruitful comments significantly improving the manuscript.
© 2023 Zoological Society of Japan
