Rediscovery and phylogenetic relationships of the scolopendromorph centipede Mimops orientalis Kraepelin, 1903 (Chilopoda): a monotypic species of Mimopidae endemic to China, for more than one century

Abstract Mimops orientalis Kraepelin, 1903 is a monotypic species of Mimopidae endemic to China. The species is known only from a single specimen, the holotype. Little is known about its biology, habitat associations, or phylogenetic relationships. It was rediscovered on Qinling Mountain in Shaanxi and Henan provinces, China, 117 years after its last record. Detailed descriptions and colour photographs of living specimens are provided along with its ecology, updated conservation notes, and data on sexual dimorphism. A genetic analysis (COI, 16S rRNA, and 28S rRNA) was conducted to assess the phylogenetic relationships among Mimopidae, Cryptopidae, Scolopendridae, Scolopocryptopidae, and Plutoniumidae. The results support classifying Mimopidae as a valid family.


Introduction
Mimops Kraepelin, 1903 is one of the least-known genera of scolopendromorph centipedes. The type species, Mimops orientalis Kraepelin, 1903, described from a single specimen from Shaanxi Province, China, has a single white ocellus on each side of the cephalic plate. Since the original description, the exact type locality of this species has been unknown, and no additional findings of M. orientalis have been reported. The genus Mimops was initially placed in the Cryptopidae and subsequently redescribed by Lewis (2006) from the holotype of M. orientalis, establishing a new monotypic family, Mimopidae Lewis, 2006. However, the only material of M. orientalis, the unsexed holotype specimen, did not allow Lewis (2006) to clarify the behaviour or information on the distribution, ecology, habitat, or phylogenetic relationships of Mimops. Because M. orientalis possesses characteristics from the families Plutoniumidae, Cryptopidae, Scolopocryptopidae, and Scolopendridae (e.g., forcipules and coxosternal plates without teeth as in Scolopocryptopidae, ultimate legs with numerous small spines as in Cryptopidae, and unpigmented ocellus-like pale areas as in Plutoniumidae), the taxonomic status of Mimopidae is still confusing, as are the phylogenetic relationships among Mimopidae, Cryptopidae, Scolopendridae, Scolopocryptopidae, and Plutoniumidae.
Throughout 2018 and 2019, we made intensive biodiversity surveys in the Taiping National Forest Park and Longyuwan National Forest Park, China. During this fieldwork, fresh material of M. orientalis was collected at the edge of Qinling Mountain. It is a region that has only few previous surveys for centipedes but has been highlighted as one of 16 biodiversity hotspots (Huang et al. 2016). Our rediscovery of this species has allowed us to reassess its taxonomic status, as well as that of the family Mimopidae, and to give fuller details of this species' morphological variations and live coloration. In this paper, we redescribe the morphological characteristics of M. orientalis and construct a phylogeny based on COI and 16S mitochondrial DNA fragments and 28S nuclear DNA fragments.

Taxonomic sampling
The studied individuals of Mimops orientalis were collected in Taiping National Forest Park, Shaanxi, China, and Funiushan National Nature Reserve, Henan, China. Other representative samples of Cryptopidae, Scolopendridae, Scolopocryptopidae, and Plutoniumidae were collected more generally in China. During the field study, some individuals were photographed with an Olympus E-M10 II camera to record their living colour pattern. The colour description is based on that of living centipedes. Specimens are deposited in the Institute of Chinese Materia Medica, China Academy of Chinese Medical Sciences, China (CMMI).
The taxonomic characteristics of the specimens were observed under an Olympus SZ16 stereomicroscope. Multifocus montage images were produced using Helicon Focus 6.7.1 software from a series of source images taken by a Canon 50D digital camera attached to the stereomicroscope. Terminology applied to the external anatomy follows Bonato et al. (2010).
All newly collected centipede specimens were stored in 75% ethanol. Legs 15 and 17 of the newly collected specimens were removed and stored in 100% ethanol for DNA extraction.

Phylogenetic sampling and DNA sequence analyses
Genomic DNA was extracted from leg tissue using a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany). Two fragments of mitochondrial DNA (mtDNA), the encoded parts of the COI and 16S ribosomal RNA genes, and a fragment of 28S nuclear ribosomal RNA were amplified according to previously literature (Vahtera et al. 2012(Vahtera et al. , 2013Siriwut et al. 2016). PCR products were bidirectionally sequenced using an ABI 3730 DNA analyser (Applied Biosystems, CA, USA). New sequences determined in this study were aligned using ClustalW in BioEdit with default settings.
Maximum likelihood and Bayesian inference approaches were employed using MEGA X (Kumar et al. 2018) and PhyloSuite (Zhang et al. 2018), respectively, to determine the phylogenetic positions of M. orientalis base on the new sequences. Standard statistical tests were applied to evaluate branch support (bootstrap support and posterior probability). The best-fit substitution models and partitioning strategies were inferred for these three datasets using ModelFinder (Kalyaanamoorthy et al. 2017) integrated into PhyloSuite (Zhang et al. 2020). Based on Akaike Information Criterion, GTR+I+G4 was chosen as the best-fitting model for ML analyses and GTR+G for Mr-Bayes analyses. Individual gene trees were constructed in MEGA using the maximum likelihood algorithm with 500 bootstrap replicates. Bayesian analyses were performed using MrBayes 3.2.6 (Ronquist et al. 2012) integrated into PhyloSuite with the default settings, and 10,000,000 MCMC generations, sampling every 100 generations. Sequences were deposited in GenBank with accessions of MT093838-MT093846, MT084401-MT084409, and MT084368-MT084375. Kraepelin, 1903 Type material. Holotype: China, Süd Schensi, August 1903, kept  Distribution. Currently, we have confirmed that M. orientalis occurs on Qinling Mountain (Shaanxi and Henan provinces) and inhabits forests near rivers above 500-1200 m elevation (Fig. 1).

Mimops orientalis
Ecology. Specimens of M. orientalis were collected from two locations on Qinling Mountain. We captured all of our new material at midday (11:00-13:00), always in forest edges near a river and under microhabitat refuges (rocks, bushes, leaf litter, and garbage). Two adult specimens were hiding under stones coated with moss, and another adult was encountered under some garbage containing rainwater. Most juveniles were encountered under small stones in bushes or moving over leaf litter in a fragmented patch of Liquidambar formosana Hance. Natural vegetation in the surrounding areas is deciduous forest, composed mainly of Juglans regia L., Cotinus coggygria Scop., and L. formosana Hance. Vegetative cover was ornamented by Urtica spp., Viola spp., Cyperus iria L., and Oplismenus undulatifolius (Arduino) Beauv (Figs 2, 3). We fed captive centipedes with mealworms, fish moths, and cockroaches, the latter two typically occurring in the centipedes' microhabitat. The centipedes became active at approximately 20:00 most nights and would quickly escape to shelter when disturbed by noise. Live colouration. Cephalic plate brownish red to orange-red, more deeply coloured in the anterior part. Coxosternites, forcipules, and tergites pale yellow to pale brown, ultimate segment and ultimate legs pale orange. Antennae, all sternites, and dorsal aspects of the legs light yellow. The live colour of a specimen from Shaanxi is slightly different from others. Its ultimate leg prefemur was nearly brown, which is more deeply coloured than that of the femur and tibia. The juvenile individual was uniformly light yellow on the cephalic plate, antennae, tergites, sternites, and all legs (Figs 4,5).
Redescription. Lewis (2006) examined and described the holotype of M. orientalis. The majority of features of the newly collected specimens are in agreement with the holotype, but with some differences.
Adult length 42-56 mm. Cephalic plate smooth, about as long as it is wide, very finely punctate with the posterior margin overlying tergite 1. Cephalic plate lacks paramedian sulci in adults (two sulci in juvenile). A pale area instead of lateral ocelli at the base of each antenna (Lewis [2006] stated that is a single ocellus, whereas Kraepelin [1903] treated this as an eyespot). The animals show no response when the pale areas are illuminated with searchlights. Antenna extend to the posterior end of tergite 8, usually with 18 articles (holotype with 17 articles on the right antenna). Articles approximately 1.20-1.42 times as long as they are wide (based on article 4). Amount of hair on the antenna gradually increases, with the presence of glabrous articles variable (Lewis [2006] stated that the basal six articles were glabrous, while Kraepelin (1903) stated that the basal seven articles were glabrous). In our collected specimens, basal articles 4 and 5 were dorsally nearly glabrous (Figs 6-9), and articles 5-7 (Fig. 8, speci- Tooth plates are short but wide, ridged and have approximately 25 ridges on each side and a prominent seta behind the anterior margin (Fig. 10). Forcipular trochanteroprefemur processes are very short, apex truncated, and with a long and acute process spine. Forcipular medial tibia and femur also have a prominent acute spine. Articles 1-3 of the second maxillary telopodite show hair and apical claw form a spine with long comb hair (Fig. 10).
All legs with two-segmented tarsi. Two tibial spurs on legs 1-18, 19 with one, 20 without. Legs 1-20 each with two tarsal spurs (Fig. 14). Legs 1-5 with one outside femur spur dorsally, legs 1-3 also with one prefemur spur dorsally (Fig. 15). Legs 1-14 with very few spines ventrally on the prefemur and femur and a distal transverse row dorsomedially on the prefemur, femur, and tibia (Fig. 14). Legs 15 and 16 with few spines ventrally on the prefemur and sparse spines ventrally on the femur. Legs 17-19 with prefemur, femur, and tibia thickly spined dorsally, medially, and ventrally, the spines on prefemur gradually increase from legs 17-19. Leg 20 with prefemur and femur thickly spined on all surfaces, tibia on all but the medial surface and tarsus 1 spined dorsally and medially.
Tergite 21 with small spines and a narrow posterior median depression (Fig. 17). Sternite 21 with sides converging posteriorly, without depression. Central zone of sternite 21 with sporadic small spines dorsal and ventral and peripheral zone with two to four rows of small spines (Fig. 16, 19). Coxopleuron with an oval pore field of many small pores and small scattered spines (Fig. 19). The coxopleural process was moderately long and digitiform, with small spines (Figs 19, 20). The prefemur and femur of the ultimate legs (Fig. 18) are covered with numerous small spines dorsally, medially, and ventrally, without grooves or strips (Lewis [2006] stated with the presence of a strip on the median ventral). Tibia spined on all but the median ventral and medial surfaces, and tarsus 1 has a few dorsomedial spines. Pretarsal accessory spurs are absent. Genital segments well developed, reaching the distance between the posterior margin of sternite 21 and the distal part of the coxopleural process. In males, sternite of genital segment 1 round, with short setae ventrally, genital segment 2 round and convex, also with short setae ventrally. Penis columnar, with long setae dorsally. Gonopod present in males with seven or eight long setae (Fig. 19). Anal valve well developed, composite by two hemispheres, with a yellow strip near the posterior margin. Genital segments short in female, posterior margin of sternite of genital segment 1 with several spines. Anal valve composed of two hemispheres, with a cavity in the centre.
Conservation status. Although this species was found from two localities separated by 400 km, there is not yet enough information about the distribution, abundance, or threats to this species, and so further surveys are needed; we consider it Data Deficient for now (IUCN 2020).
The results of the phylogenetic analysis are presented in Figures 21 and 22. Bayesian inference (BI) and maximum likelihood (ML) analyses yielded trees based on the combined alignments of COI + 16S + 28S, which demonstrated essentially consistent topologies. All Scolopendromorpha species used in this study could be considered as two main clades, the ocellate scolopendromorphs and the blind scolopendromorphs, both well supported in the ML tree. Mimopidae was positioned as sister group to the other families of blind scolopendromorphs. However, Bayesian inference analyses identified Scolopendromorpha species as belonging to three clades, the ocellate scolopendromorphs, the blind scolopendromorphs sensu Pocock, 1896 (comprising Plutoniumidae, Cryptopidae, and Scolopocryptopidae), and the monophyletic family Mimopidae with quite strong support (PP 0.94). Both methods obtained from the molecular data place Scolopendridae as a monophyletic family, comprising two subfamilies, Scolopendrinae and Otostigminae. Likewise, Scolopocryptopidae comprises two subfamilies, Scolopocryptopinae and Newportiinae. Cryptopidae, and Scolopocryptopidae are sister groups together with Plutoniumidae in the molecular analysis.

Discussion
The taxonomic position of Mimops has been controversial for many years. Attems (1930) divided scolopendromorphs into blind scolopendromorphs and ocellate scolopendromorphs, the latter containing only a single family, Scolopendridae, whereas the former includes three eye-less scolopendromorph families (Cryptopidae, Plutoniumidae, and Scolopocryptopidae), hence, the name blind clade. The genus Mimops was first classified into the subfamily Cryptopinae of the family Cryptopidae by Attems (1930) and then removed from Cryptopinae in a cladistic analysis by Schileyko and Pavlinov (1997). Shelley (2002) moved Mimops to the Scolopendridae, and subsequently Lewis (2006)  Mimops also shares many scolopendrid-like characteristics, but distinct differences from Tonkinodentus and Cormocephalus imply that eye loss may have occurred several times in the evolution of Scolopendromorpha. Mimops orientalis also possesses several unusual characteristics among Plutoniumidae, Cryptopidae, Scolopocryptopidae, and Scolopendridae. It shares many characteristics with scolopendrid centipedes, including a smooth cephalic plate without setae or spines, several glabrous basal antennal articles, the presence of sternite and tergite paramedian sutures, the presence of spines on the prefemora of the ultimate legs, and an anterior transverse sulcus on tergite 1. These characteristics render it similar to Scolopendra species, particularly many New World varieties. However, M. orientalis lacks lateral ocelli and instead possesses a pale area, which is only present in plutoniumids, and the spines on the ultimate legs are very small, which is similar to some Trigonocryptops species, e.g., Cryptops (Trigonocryptops) camoowealensis Edgecombe, 2006. This renders the taxonomic position of Mimops puzzling.
In this study, we confirmed Mimopidae as a valid family based on both morphological and molecular data. We also place Mimops as the basal taxa of scolopendromorphs for the following reasons: 1) M. orientalis is resolved as monophyletic both in maximum likelihood and Bayesian inference analyses under all parameter sets that were explored; 2) the ocelli of Mimops are regressed and have left two pale areas, whereas the outgroups Lithobiomorpha, Scutigeromorpha, and Craterostigmomorpha all have ocelli (or a compound eye with ommatidia in the case of Scutigeromorpha), indicating that ocelli may be a plesiomorphic characteristic; 3) Mimops bears 21 trunk segments, whereas 21 to 23 segments is an unreversed apomorphy of Scolopocryptopidae; 4) the long and acute spines on the forcipular trochanteroprefemoras of Mimops may be a primitive characteristic of trochanteroprefemur process; and 5) the small homologous spines on the ultimate legs, coxopleural process and ultimate sternites seem to be an intermediate state between setae and spines. However, although BI and ML analyses both support the validity of Mimopidae, the precise position of Mimops is not delimited due to the unresolved trichotomy in BI tree. Further research based on complete mitochondrial genome or transcriptomic data may help to clarify phylogenetic relationships among Mimopidae and other families.