Integrative taxonomy of Dicellophilus Cook, 1896 (Chilopoda, Geophilomorpha, Mecistocephalidae) in Japan, with a description of a new species

The genus Dicellophilus Cook, 1896, is a peculiar genus from the point of view of distribution. Dicellophilus is distributed in three limited areas that are well separated from one another: central Europe ( D. carniolensis ), Honshu ( D. pulcher ), and the southwestern part of the USA ( D. anomalus and D. limatus ). In the present study, in a field survey conducted throughout Japan, specimens belonging to the genus Dicellophilus were collected from Tohoku to the Kansai region, Honshu. Morphological analysis, molecular phylogenetic analysis, and genetic distance among Dicellophilus in Japan and D. carniolensis revealed that specimens from Sendai-shi, Miyagi Pref., could be assigned to an undescribed species. This previously unrecognized species is herein described as D. praetermissus sp. nov. The new species can be distinguished from D. carniolensis and D. limatus by the number of pairs of legs (43 pairs in D. carniolensis and 45 in D. limatus , but 41 in D. praetermissus sp. nov. ), from D. anomalus by the lack of a pair of setae on the posteromedian part of the clypeus and variable crenulation on the internal margin of the forcipular tarsungulum, and from D. pulcher based on the following combination of characteristics: both ends of the transverse suture not evidently convex forward; long rather than wide trochanteroprefemur; wide rather than long metasternite.


Introduction
The geophilomorph family Mecistocephalidae Bollman, 1893, is mainly distributed from temperate to tropical regions in both hemispheres, and species diversity is remarkably high in Japan (Uliana et al. 2007;Bonato 2011).To date, approximately 180 species are known worldwide, and 31 species have been recorded from Japan (Uliana et al. 2007;Tsukamoto et al. 2019Tsukamoto et al. , 2022)).Therefore, approximately 20% of all known mecistocephalid species are distributed in Japan.Moreover, Japan is the richest with regard to the number of genera (nine out of 11 genera; Uliana et al. 2007;Tsukamoto et al. 2022).
Mecistocephalidae are morphologically characterized by a cephalic capsule and a forcipular segment that are evidently sclerotized and darker than the remaining trunk segments (Bonato et al. 2003(Bonato et al. , 2014;;Uliana et al. 2007;Bonato 2011).In addition, the following three features characterize Mecistocephalidae: a mandible with a series of pectinate lamellae only; trunk sternites with an internal apodeme; and a mid-longitudinal sulcus (Bonato et al. 2003).Notably, the segment number of most species of Mecistocephalidae has no intraspecific variation, except for some species of the genus Mecistocephalus Newport, 1843, with a very high number of leg-bearing segments (Bonato et al. 2003(Bonato et al. , 2014;;Uliana et al. 2007;Bonato 2011).
The distribution of Dicellophilus is limited to three areas that are well separated from one another: central Europe (D.carniolensis), Honshu of Japan (D. pulcher), and the southwestern part of the USA (D. anomalus and D. limatus) (Bonato et al. 2003(Bonato et al. , 2010a)).All three distribution areas are located within a narrow latitudinal band, at approximately 35-45°N, and all species inhabit humid litter and soil in forests (Bonato et al. 2010a).Bonato et al. (2010a) performed phylogenetic analyses of the genus Dicellophilus with a morphological dataset consisting of 30 characteristics and indicated that the two American species, viz., D. anomalus and D. limatus, were sister species, and D. pulcher formed a clade with the two American species.As far as extant species are concerned, D. carniolensis is thus the first diverged species in the genus Dicellophilus.
For a decade, the combination of morphological observation, molecular phylogenetic analyses, and DNA barcoding ("integrative taxonomy") has helped detect undescribed species and reveal the genetic structure of the taxa concerned (Dayrat 2005;Padial et al. 2010).For example, some studies using integrative approaches to scolopendromorphs and geophilomorphs have revealed the existence of many cryptic species under one validly named species or distinct morphospecies (Joshi and Karanth 2012;Siriwut et al. 2015Siriwut et al. , 2016;;Tsukamoto et al. 2021b;Peretti et al. 2022;Bonato et al. 2023).Tsukamoto et al. (2022) also detected the existence of two species of the mecistocephalid genus Nannarrup Foddai, Bonato, Pereira & Minelli, 2003 by using an integrative approach.
Inspired by these previous studies, the present study aims to reveal the genetic diversity and confirm whether the morphospecies D. pulcher involves unnoticed and undescribed cryptic species by using an integrative taxonomic approach.

Taxon sampling
Although our ongoing sampling efforts to taxonomically reveal the East Asian mecistocephalid faunas cover the whole of Japan and surrounding areas, the present study focused on Honshu (the largest island of mainland Japan), from which D. pulcher was described.
A total of 38 specimens morphologically identified as Dicellophilus pulcher, hitherto the only known Japanese species of the genus, were collected from Honshu from 2018 to 2021.The detailed collection sites of the examined specimens are shown in Fig. 1, Table 1, and the "Taxonomic account" section.The altitude data provided by AW3D of JAXA (https://www.eorc.jaxa.jp/ALOS/jp/index_j.htm)and the coastal line provided by the digital nation land information (https://nlftp.mlit.go.jp/index.html) were used to generate Fig. 1.
Each specimen was labeled with its unique specimen identification number in the form "TSYYYYM-MDD-XX," where TS is an abbreviation of the first author's name, Tsukamoto Sho; YYYYMMDD designates the date on which the specimen was collected; and XX is the identification number assigned to each specimen collected on a particular date (e.g., TS20171010-01).
All of the type specimens of Dicellophilus designated in this paper were deposited at the Collection of Myriapoda, Department of Zoology, National Museum of Nature and Science, Tokyo (NSMT), and the Museum of Nature and Human Activities, Hyogo (MNHAH).The deposition site of each type specimen is shown in the "Taxonomic account" section.All non-type voucher specimens of Dicellophilus pulcher are managed by the first author.

Morphological examination
For dissected specimens, the cephalic capsule, maxillae, mandibles, forcipular segment, and leg-bearing segments were made transparent using lactic acid to examine the anatomy and produce images.Multi-focused images of these body parts were produced using Affinity Photo 1.10.4(https://affinity.serif.com/ja-jp/photo/)from a series of source images taken using a Canon EOS Kiss X9 digital camera attached to a Nikon AZ100 microscope and improved using Adobe Photoshop Elements 10 and Affinity Designer 1.10.5 (https://affinity.serif.com/ja-jp/designer/).Then, the body parts were measured directly using an ocular micrometer attached to the microscope.
The morphological terminology used in this study is in accordance with Bonato et al. (2010b).Specimens with fully developed paired gonopods, that is, evidently biarticulated in males and touching one another in females, were determined to be adults, and those with incompletely developed paired gonopods were determined to be subadults.Specimens without gonopods were determined to be juveniles based on Uliana et al. (2007).In the present study, 23 adult specimens out of 38 collected were examined morphologically.

DNA sequencing
Genomic DNA was extracted from one or two legs of each specimen in accordance with the Chelex-TE-ProK protocol described by Satria et al. (2015), with incubation for 4-24 h.  2. Each PCR product was screened by electrophoresis on a 2.0% agarose gel in 1× TAE.
The amplification conditions for mitochondrial COI were as follows: 98 °C for 2 min; 5 cycles of 98 °C for 10 s; 45 °C for 30 s; and 68 °C for 45 s; 40 cycles of 98 °C for 10 s; 48.5 °C for 30 s (annealing step); 68 °C for 45 s; and 68 °C for 7 min.If the target fragment of COI was not appropriately amplified, then the annealing temperature was changed from 48.5 °C to 50 °C.PCR was performed again by omitting the first five cycles of annealing and the extension step.
The amplification conditions for mitochondrial 16S were as follows: 98 °C for 2 min; 35 cycles of 98 °C for 10 s; 45 °C for 30 s (annealing step); 68 °C for 45 s; and 68 °C for 7 min.If the target fragment of 16S was not appropriately amplified, then the annealing temperature was changed from 45 °C to 48 °C.The number of annealing cycles was changed from 35 to 45.
The amplification conditions for nuclear 28S were as follows: 98 °C for 2 min; 5 cycles of 98 °C for 10 s; 42 °C for 30 s; and 68 °C for 1 min; 30 cycles of 98 °C for 10 s; 50 °C for 30 s (annealing step); 68 °C for 1 min; and 68 °C for 7 min.If the target fragment of 28S was not appropriately amplified, then the annealing temperature was changed from 50 °C to 48 °C.The number of annealing cycles was changed from 30 to 40-45 cycles.In addition, PCR was performed again by omitting the first five cycles of annealing and the extension step.
The amplified products were incubated at 37 °C for 4 min and at 80 °C for 1 min using ExoSAP-IT TM Express (Thermo Fisher Scientific) to remove any excess primers and nucleotides.All nucleotide sequences were determined by direct sequencing using the ABI PRISM BigDye TM Terminator Cycle Sequencing Kit ver.3.1 (Thermo Fisher Scientific) or BrilliantDye TM Terminator Cycle Sequencing Kit v. 3.1 (Nimagen, B.V., Nijmegen, Netherlands) equipped with an ABI 3130xl automated sequencer (Thermo Fisher Scientific).The sequences were assembled using ChromasPro 1.7.6 (Technelysium Pty Ltd., Australia) and deposited onto the DDBJ, EMBL, and GenBank databases under the accession numbers LC815125-LC815233 (Table 1).

Molecular phylogenetic analyses
The sequences obtained using the abovementioned methods were used for phylogenetic analyses, together with the COI, 16S, and 28S sequences of Dicellophilus carniolensis (C.L. Koch, 1847) and Nannarrup innuptus Tsukamoto in Tsukamoto et al. (2022) obtained from GenBank as outgroups (Table 1).The datasets for COI (658 bp positions), 16S (514 bp positions), and 28S (971 bp positions) were concatenated to form the COI + 16S + 28S dataset for phylogenetic analyses.All sequences were aligned using MAFFT v. 7.475 (Katoh and Standley 2013).For COI, alignment was performed using the default setting.For 16S and 28S, secondary structure alignment was performed using the X-INS-i option.
Maximum-likelihood (ML) trees were created on the basis of the sequence dataset for each gene and concatenated using IQ-tree 1.6.12(Nguyen et al. 2015).As an optimal substitution model in accordance with BIC, TNe + I + G4 was selected for the first codon position of COI in the concatenated dataset; TNe + G4 for the COI dataset; HKY + F was selected for the second codon position of COI in both datasets; TN + F + G4 was selected for the third codon position of COI in both datasets; HKY + F + I + G4 was selected for 16S of both datasets; and TIM3e + G4 was selected for 28S of both datasets.Ultrafast bootstrap analysis (UFBoot; Hoang et al. 2018) and the SH-like approximate likelihood ratio test (SH-aLRT; Guindon et al. 2010) were performed with 1,000 replicates.
Bayesian inference trees were created using ExaBayes 1.4.1 (Aberer et al. 2014) under the default substitution model "GTR + G." The Markov chain Monte Carlo method was used with random starting trees and performed once for each of the four chains (three hot and one cold) for 10,000,000 generations for each dataset except COI, but for 20,000,000 generations for the COI dataset.Trees were sampled every 500 generations, tuning parameters every 100 generations, and the first 25% of the trees were discarded as burn-in.Other parameters were set in accordance with the default settings.The effective sampling size of each parameter was confirmed to be 200 using Tracer 1.7.1 (Rambaut et al. 2018).

Calculation of the genetic distances
The aligned COI dataset used for phylogenetic analyses were also used to calculate genetic distances.Kimura two-parameter (K2P) distances were calculated using MEGA X (Kumar et al. 2018) with the setting "pairwise deletion."

Delimitation of "provisional" operational taxonomic units
The program "assemble species by automatic partitioning (ASAP)" was used to delimit "provisional" operational taxonomic units (POTUs).ASAP is a species delimitation program based on a hierarchical clustering algorithm that only uses pairwise genetic distances (Puillandre et al. 2021; available at https://bioinfo.mnhn.fr/abi/public/asap/).ASAP was performed for the COI sequence dataset of Dicellophilus (excluding the outgroup) under the "pairwise K2P distance" method.

Delimitation of putative species and provisional naming of each putative species
The present study preliminarily relied on the morphological information provided by Bonato et al. (2010a) as the basis for the monophyly of the genus Dicellophilus.Then, putative species were proposed.Except for the assumption of monophyly of the abovementioned genus, the following steps generally followed the workflow "DI-system" proposed by the first author in Tsukamoto (2023) for discriminating and labeling putative species: (I) sorting specimens, which are morphologically conferrable to Dicellophilus, into morphospecies; (II) confirming the monophyly of the morphospecies with the phylogenetic tree inferred by the sequence dataset of three gene markers, mitochondrial COI and 16S, and nuclear 28S; (III) calculating the genetic distance of COI between congeneric morphospecies, which are confirmed to be monophyletic, to define the intermorphospecific threshold; (IV) delimiting POTUs using ASAP (see above) and confirming the most conferrable hypotheses of species-level independence by considering the phylogenetic tree and the intermorphospecific threshold defined in step III.By steps III and IV, species hypotheses can be established from two viewpoints, viz., phylogeny and clustering based on DNA data.In step IV, putative species were recognized by considering three species delimitation principles: (1) each clade is regarded as an independent putative species if it diverges from all others with a minimum K2P distance higher than the intermorphospecific threshold; (2) a single putative species that satisfies (1) can contain inner lineage(s) diverging extremely from the others unless the maximum distance from the sister inner lineage exceeds the intermorphospecific threshold; (3) a single putative species that satisfies (2) cannot contain inner lineage(s) diverging extremely from the sister inner lineage with the minimum K2P distance exceeding the intermorphospecific threshold, and such a lineage must be further considered as a distinct species if it exists.
As mentioned above, the present study presupposes the monophyly of Dicellophilus, supported by the morphological evidence (Bonato et al. 2010a).This is because the possibility of a difference in evolutionary rate among genera is important for defining the intermorphospecific threshold in the "DI-system"."DI-system" is planned to be proposed formally in future studies.CCT TGG TCC GTG TTT CAA GAC Edgecombe and Giribet (2006) Each putative species recognized by following the abovementioned steps was also labeled in accordance with the study of Tsukamoto ( 2023), with a unique, permanent, and citable identifier "DI," such as "0000-0003-3020-8454_XXXX," in which "0000-0003-3020-8454" shows the author's ORCID and "XXXX" shows a unique identification number given to each species in the author's life-long research.ORCIDs involved in the species identification codes were omitted except for section titles, figure legends, and tables to avoid redundancy.
Depending on the availability of the morphological information necessary to formally describe and name species in the conventional manner of Linnaean Taxonomy (The International Commission on Zoological Nomenclature 1999), the putative species labeled with the DI can be described and named (step V).In the present study, species discrimination (steps I-IV) and formal description and naming of the species (step V) were separated as two methodologically distinct phases.
The ML tree based on the 16S dataset (Fig. 4) and the BI tree based on the same dataset (only PP shown in Fig. 4) involving 38 specimens also show Clade A (UF-Boot = 100%, SH-aLRT = 100%, PP = 1.00).Clade A is deeply separated from all other Dicellophilus specimens, but Clade B is not well supported (UFBoot = 78.2%,SH-aLRT = 63%, PP = 0.91).In addition, the phylogenetic relationship among Clade B and other Dicellophilus specimens was not clear due to low support values.
The ML tree based on the 28S dataset (Fig. 5) and the BI tree based on the same dataset (only PP shown in Fig. 5) involving 33 specimens also show Clade A (UFBoot = 99.7%,SH-aLRT = 100%, PP = 1.00).Clade A is deeply separated from all other Dicellophilus specimens, and Clade B conforms to a further clade with D. carniolensis, like the topology of the concatenated dataset (UFBoot = 93.6%,SH-aLRT = 97%, PP = 0.99).However, the phylogenetic relationship among Clade B and other Dicellophilus specimens was not clear due to low support values.Although there is no consistency of phylogenetic relationship among Clades A, B, and D. carniolensis in four datasets, each topology shows that Clade A is a distinct lineage from other Dicellophilus specimens.
Etymology.The species name is a masculine adjective derived from "overlooked" in Latin.Since the description by Kishida (1928) of D. pulcher (as Mecistocephalus pulcher), this new species has been overlooked for 90 years, despite documentation of its distribution as Dicellophilus in the Sendai-shi, Miyagi Pref.(Takakuwa 1940).
Diagnosis.Trunk segments without dark patches; head 1.4 times as long as wide; lateral margin of cephalic plate abruptly converged posteriorly; clypeus with densely scattered setae; palaclypeal suture evidently converging posteriorly; both ends of transverse suture uniformly rounded; mandible with 6 lamellae; forcipular trochanteroprefemur longer than wide, with one small distal denticle; forcipular tarsungulum with evident and variably spaced notches; metatergite subtrapezoidal; metasternite trapezoidal, wide rather than long; forty-one pairs of legs.
Cephalic capsule (Fig. 7A, B): Cephalic plate ca 1.3-1.4×as long as wide; lateral margins markedly convex; posterior margin straight; areolate part visible only at anterior margin; scutes approximately isometric and up to 20 μm wide in 50 mm long specimen; both ends of transverse suture uniformly rounded or slightly convex forward; setae up to ca 300 μm long.Clypeus ca 2.3-2.5× as wide as long, with lateral margins complete, anterior part areolate, with scutes ca 30 μm wide in 50 mm long specimen, clypeal areas absent; clypeus with about 200 setae on most part except lateral and posterior margins; clypeal plagulae undivided by mid-longitudinal areolate stripe.Anterior and distolateral parts of pleurites areolate, without setae, non-areolate part extending forwards distinctly beyond labrum.Side-pieces of labrum not in contact, anterior margin not concave posteriorly but horizontally, divided into anterior and posterior alae by chitinous line, with longitudinal stripes on posterior alae, with medial tooth, and short fringe on posterior margin of side-pieces; mid-piece ca 6.2 times as long as wide, lateral margin concaved.
Antenna (Fig. 8A-H): Antenna with 14 articles, when stretched, ca 2.7-3.2× as long as head length.Intermediate articles longer than wide.Distal part of article areolate, remaining surface not areolate in article I-XIII.Article XIV ca 2.1-2.5× as long as wide, ca 1.1-1.5×as long as article XIII.Setae on articles VIII-XVI denser than articles I-VII.Setae gradually shorter from article VIII to XIV, up to ca 290 μm long on article I, up to ca 270 μm long on article VIII and < 75 μm long on article XIV.Article XIV with two types of sensilla; apical sensilla (arrows in Fig. 8G, H) ca 25 μm long, with wide flat ring at mid-length; club-like (arrowheads in Fig. 8G, H) sensilla ca 15 μm long, clustered in distal part of internal and external sides of article.Rows of spine-like basal sensilla (the 'sensilla microtrichoidea' of Ernst 1983Ernst , 1997Ernst , 2000) ) absent on antennal article VI and X.A few pointed sensilla, up to 7.5 μm long, on both dorso-external and ventro-internal position, close to distal margin of articles II, V, IX and XIII.

Species identified in the present study
Both ends of transverse suture
Forcipular segment (Fig. 10A-E): Tergite trapezoidal, ca 1.3-1.4×as wide as long, with lateral margins converging anteriorly, areolation mainly along two marginal lateral and anterior bands and two paramedian posterior areas, gradually fading into central non-areolate surface; ca 0.5-0.6×as wide as cephalic plate and ca 0.4-0.5×as wide as tergite 1; 3+2 setae of similar length arranged in an anterior row, and ca 20 setae of similar length arranged symmetrically in a posterior row.Mid-longitudinal sulcus of tergite not visible.Pleurite 1.8-1.9×as long as the tergite; dorsal ridge sclerotized; anterior tip (scapular point) well behind anterior margin of coxosternite, and only slightly projecting.Cerrus composed of a group of 10-20 setae on each side of anterodorsal surface of coxosternite, but no paramedian rows of setae.Exposed part of coxosternite ca 1.2× as wide as long; anterior margin with shallow medial concavity and with one pair of denticles; coxopleural sutures complete in entire ventrum, sinuous and diverging anteriorly; chitin-lines absent; condylar processes of forcipular coxosternite well developed.Trochanteroprefemur ca 1.3-1.4×as long as wide; with a pigmented tubercle at distal internal margin.Intermediate articles distinct, with a tubercle on femur and tibia.Tarsungulum with well-pigmented basal tubercle on dorsal surface; both external and internal margins uniformly curved, except for moderate mesal basal bulge; ungulum not distinctly flattened; internal margin of ungulum evidently crenulated, with variably spaced notches.Elongated poison calyx lodged inside intermediate forcipular articles.
Male postpedal segments (Fig. 12A, B): Two gonopods, very widely separated from one another, conical in outline, bi-articulated with sutures, covered with setae.Anal pore present.
Distribution.Only known from the type locality.
Remarks.Dicellophilus praetermissus sp.nov.most closely resembles D. pulcher but is distinguishable by the following combination of characteristics: both ends of transverse suture not evidently convex forward; the longer than wide trochanteroprefemur; the wide rather than long metasternite (Table 3).

Species hypothesis for Japanese Dicellophilus
No POTU delimitation hypotheses proposed by ASAP corresponded well to the principle in steps I-IV.This is because each of those hypotheses involves many POTUs, which were separated from each other with a K2P distance lower than the intermorphospecific threshold.Although Clade B could be divided into many more POTUs, such partitioning hypotheses can be rejected as oversplitting in accordance with the species delimitation criteria in step IV with the intermorphospecific threshold (21% in COI).Therefore, two putative species were recognized in the 38 D. pulcher specimens examined in steps I-IV, and they can be separately labeled as follows: Dicellophilus sp.0000-0003-3020-8454_0068 (= Clade B; hereafter referred to as D. sp.0068) and D. sp.0000-0003-3020-8454_0069 (= Clade A; hereafter referred to as D. sp.0069, Table 3, Figs 2-5).
Only one validly named species of Dicellophilus from Japan, D. pulcher, was described on the basis of a specimen from the Subashiri trail of Mt.Fuji, Shizuoka Pref.(Kishida 1928).Dicellophilus sp.0068 involves a specimen from the type locality (TS20210504-01) and shares the number of pairs of legs, the presence of macropores on the coxopleuron, and a wide forcipular trochanteroprefemur with the original description (including figures) of D. pulcher.Therefore, D. sp.0068 is conferrable to D. pulcher.By contrast, D. sp.0069 can be regarded as a new species based on morphological comparison with other congeners, including D. pulcher.This new species is described in the "Taxonomic account" section under the name Dicellophilus praetermissus sp.nov.(step V); see also the taxonomic discussion of junior synonyms of D. pulcher.

Oversplit of the number of POTUs by ASAP
As mentioned in the result section, many POTUs were divided from the COI dataset of Dicellophilus examined in the present study.When taking into account overall genetic diversity of all examined specimens of Dicellophilus, i.e., D. pulcher, D. carniolensis and D. praetermissus sp.nov.(maximum genetic divergence in K2P: 24%), the genetic diversity within the morphospecies D. pulcher alone is quite high (15%).
Possibly, such an oversplit of POTU would have been caused by the algorithm ASAP and the quite high genetic divergence of D. pulcher in the dataset.According to Puillandre et al. (2021), ASAP is a hierarchical clustering algorithm.Each subgroup was separated depending on the average pairwise distance between subgroups and within the subgroup, sample size, and a coalescent mutation rate.Based on this algorithm, the distribution of genetic distances will affect the result of the number of species (= POTUs).In detail, when there is high genetic diversity within one morphospecies compared to the whole dataset, the morphospecies will be divided into several POTUs, in accordance with the possibility of panmixia (p-value).
POTUs are divided by the possibility of panmixia, so it can be expected that each POTU will be a biological species.However, each POTU within the morphospecies D. pulcher detected in the present study is not regarded as a species until morphological evidence is discovered.

Distribution of Dicellophilus in Japan
Dicellophilus specimens examined in Japan were collected from 34°33'N to 38°16'N on Honshu within a latitudinal band, which is congeners' distribution.The authors and their collaborators have collected Dicellophilus exclusively from Miyagi Pref. to Kyoto Pref.but not from other areas, despite a comprehensive field trip in Japan (Fig. 1).Therefore, the distribution of Dicellophilus in Japan should be restricted from Miyagi Pref. to Kyoto Pref.(see Remarks of D. pulcher in this paper;Takakuwa 1940;Uliana et al. 2007).
According to the molecular phylogenetic analyses of Dicellophilus specimens in Japan, it is possible that there are two large populations among D. pulcher, viz., specimens from Eastern Honshu (Fukushima Pref. to Shizuika Pref.and one specimen from Gifu Pref.) and those from western Honshu (Gifu Pref. to Kyoto Pref.), because the monophyly was supported by the phylogenetic analysis based on the concatenated and COI datasets, respectively.However, the boundary between two populations is still not clear due to the lack of field surveys in the central part of Honshu.
In field surveys conducted by the authors in Japan, D. praetermissus sp.nov.was collected only in Sendai-shi, Miyagi Pref.(the northern part of Honshu).It is also noteworthy that D. pulcher has yet to be collected from the northern part of Honshu (from Aomori Pref. to Miyagi Pref).This result of field surveys shows that the distribution of D. praetermissus sp.nov.may segregate from D. pulcher, but further field surveys are needed around Miyagi Pref.

Figure 1 .
Figure 1.Map of the collection sites of specimens examined in the present study.White circle, Dicellophilus pulcher; black circle, D. praetermissus sp.nov.

Figure 2 .
Figure 2. Maximum-likelihood tree of Dicellophilus based on the concatenated dataset of COI, 16S, and 28S, with the results of species delimitation by ASAP.Note that specimens whose COI sequence was not determined were included in the conferred species if they belonged to the same concerning clade to easily understand the result of species delimitation.Nodal values are obtained from the ultrafast bootstrap (UFBoot), SH-like approximate likelihood ratio test (SH-aLRT), and posterior probability (PP).The asterisk (*) indicates 100% in UFBoot, SH-aLRT, and 1.0 in PP.Hyphen (-) indicates lower than 95% in UFBoot, 80% in SH-aLRT, or 0.95 in PP.Nodal values are not shown when UFBoot, SH-aLRT, and PP values are <95%, <80%, and <0.95, respectively.The unit of evolutionary distance is the number of base substitutions per site.A broken square shows that the clade consisted of specimens from eastern Honshu.Abbreviations: Ai = Aichi Pref.; Fs = Fukushima Pref.; Gi = Gifu Pref.; Kn = Kanagawa Pref.; Ky = Kyoto Pref.; Mi = Mie Pref.; My = Miyagi Pref.; Ni = Niigata Pref.; Nn = Nagano Pref.; Sh = Shizuoka Pref.; Si = Saitama Pref.;To = Tokyo Pref.; Yn = Yamanashi Pref.
The best five partitioning hypotheses inferred by the ASAP program are shown in Figs 2, 3, with the following ASAP scores: (1) 27 POTUs with a score of 3.0; (2) 18 POTUs with a score of 4.0; (3) 19 POTUs with a score of 4.0; (4) 28 POTUs with a score of 6.0; and (5) 24 POTUs with a score of 7.0.Morphological examination of Japanese DicellophilusAll 38 specimens of D. pulcher (a combination of Clades A and B) examined in steps I-IV have 41 pairs of legs and can be distinguished from D. carniolensis and D. limatus by the number of pairs of legs (43 pairs in D. carniolensis and 45 in D. limatus).Examined 23 adult specimens can also be distinguished from D. anomalus, which has 41 pairs of legs, by the lack of a pair of setae on the posteromedian part of the clypeus and variable crenulation on the internal margin of the forcipular tarsungulum(Bonato et al. 2010a).
New Japanese name: Date-hirozujimukade

Table 1 .
The list of specimens that were used in the phylogenetic analyses.Geographic coordinates enclosed by parentheses are secondary due to the lack of information in the labels.

Table 2 .
The list of primers used in the present study.