Three new and one little-known species of Hypogastruridae (Collembola) from Russia’s northeast

Abstract Three new species, Hypogastrura variatasp. nov., Xenylla aculeatasp. nov., and X. arneisp. nov., are described based on material from coastal communities of the Sea of Okhotsk, northeast Russia. Taxonomic remarks concerning a little-known species, H. yosii Stach, 1964, found in coastal wrack sediments on Kunashir Island, Kuriles are also given. COI sequences of the above species are analysed, thus allowing for their species statuses to be confirmed.


Introduction
The first information concerning the collembolan faunas of the Far Eastern regions of Russia was obtained during the expedition of N.A.E. Nordenskjöld aboard the 'Vega' along the northern coasts of Eurasia to the Commander Islands (1878-1879). The material collected by that expedition on the northern coast of Chukotka was published by Schött (1893), who recorded slightly fewer than 20 species of Collembola from the region. Currently, this number totals 466 species (Kuprin and Potapov 2018) and continues to increase. This paper is devoted to descriptions of three new species of the family Hypogastruridae from the coastal communities in the northern part of the Sea of Okhotsk. Moreover, their statuses are confirmed using both morphological and molecular studies.
lateral process on the labial palp;

Methods of molecular analysis
DNA was isolated from specimens fixed in 96% ethanol using Holterman's technique (Holterman et al. 2006), with the addition of proteinase K and mercaptoethanol in the lysing solution. Sequences of the cytochrome oxidase subunit I (COI) gene were amplified using an EncycloPlus PCR Kit (Evrogen, Russia) with primers from the Table 1. The standard PCR reaction protocol of the Canadian Center for DNA Barcoding was used for amplifications (http://www.dnabarcodes2011.org/conference/ preconference/CCDB-Amplification-animals.pdf ). Polymerase chain reaction (PCR) products were visualised in gel, cut out, and cleaned using the SV Gel and PCR CleanUp System kit (Evrogen, Russia). DNA sequencing was performed at the Genome Centre for Collective Use in the Severtsov Institute of Ecology and Evolution of Russian Academy of Science (Moscow, Russia). The sequences were combined and aligned using ClustalX software after the addition of sequences from the GenBank. Distance analyses were performed with MEGA6 (Tamura et al. 2013) with the Kimura-2 parameter model (Kimura 1980) to estimate genetic distances. All sequences were deposited into the GenBank. Additional material. more than 300 specimens (alcohol), mainly from the P. phryganodes plant association of the same region. Several specimens from this material were sequenced (Table 1). Their partial COI genes were amplified and deposited in the GenBank under the sample ID: KY066780-KY066783.

Species descriptions
Diagnosis. A species of the genus Hypogastrura Bourlet, 1839, with four weakly differentiated, curved, sensory setae (one dorsal and three external) on Ant.4, the relatively clearly differentiated dorsal setae, the tridentate retinaculum, the basal lamella on the unguiculus, one tenent seta on all tibiotarsi, the partly reduced furca with four or five posterior setae, and the highly variable shape of the mucro.
Variability. One of the most characteristic features of the new species is its highlevel variability of such important morphological traits as the number of PAO lobes and dental setae, as well as the shape and presence of a mucro (Table 2). This may be assumed as a direct consequence of rather severe conditions of boreal maritime marshes. Etymology. The name of the new species is intended to reflect the morphological variability.
Affinities. Apart from the new species, there are only four congeners known in the world fauna that are characterised by a shortened furca with five or fewer dental setae, coupled with a tridentate retinaculum and only one tenent seta on each leg: H. oreophila Butschek, 1948, H. exigua Gisin, 1958 (Nosek, 1976), and H. magistri Babenko, 1994. The first two species are from high-montane habitats in the European Alps, and both have been recently redescribed (Skarżyński 2011). They are much smaller than H. variata sp. nov. (0.8 mm vs. 1.2-1.8 mm) and have short, undifferentiated, dorsal setae and a longer furca (dens+mucro/U 3 with ratio ~ 2 vs. 1.0-1.3 in H. variata sp. nov.). In addition, H. oreophila is characterised by the presence of m-setae on Abd.5, a broadened maxillary L.1, an inner tooth on the unguis and a hook-like mucro with a broad outer lamella, whereas H. exigua shows a larger PAO (~1.5 ocellus) and more numerous setae on VT (5+5 vs. 4+4 in H. variata sp. nov.).
The two other similar species, H. mongolica and H. magistri, are known from mountainous regions of Central Asia (northern Mongolia and western Tuva). Of these, H. magistri can easily be distinguished due to the presence of six or seven curved sensilla on Ant.4 (vs. four in H. variata sp. nov.) and the presence of additional setae on Abd.4 and Abd.5. As regards H. mongolica, its comparison with H. variata sp. nov. is impossible, because the holotype, the only known specimen, of H. mongolica was immature (Skarżyński 2011). According to the original description (Nosek 1976) and redescription of the type (Skarżyński 2011), H. mongolica differs from H. variata sp. nov. in being smaller (0.6 mm long) and lighter in colouration, the "body clothed sparsely with short setae", long tergal sensilla and an inner tooth on the unguis, but at least some of these characters may reflect its immature status. Nevertheless, H. mongolica and H. variata sp. nov. are unlikely to be synonymous from an ecological point of view alone, because their habitat preferences are drastically different: litter of a mountain forest vs. a saline maritime marsh.
There are another four known Palaearctic congeners which may be related to the above group: H. capitata Cassagnau & Delamare, 1955 (Lebanon), H. verruculata Rusek, 1967 (China), H. ramia Lee & Choe, 1979 (South Korea), and H. pizzoci Fanciulli & Dallai, 2008 (Italy). All of them are also characterised by the presence of a single tenent seta on each leg, and the unguiculus with a basal lamella and a tridentate retinaculum, but they all share a complete, less strongly reduced furca with six posterior setae.
Molecular data. Unfortunately, the GenBank does not contain sequences of any of the above-mentioned species. Therefore, the isolated position of H. variata sp. nov. among fifteen taxonomic units of Hypogastrura present in the GenBank is not particularly surprising and may well serve as an additional confirmation of its independent status. Molecular data have shown that the divergences between all units considered are rather high (Table 3). The average interspecific divergence between all species is 26.3% (ranging from 15.9 to 36.4%), while it is 26.7% for H. variata sp. nov. and the other fourteen species (ranging from 23.0 to 31.4%). Nevertheless, it seems noteworthy that the molecular trees obtained fail to fully reflect the relationships within the genus Hypogastrura based on morphological evidence alone. The most apparent assumption explaining this fact is that the molecular data are still too scant to realistically construct reliable trees that would adequately reflect the real phylogenetic relationships within the genus.
Distribution and ecology. Hypogastrura variata sp. nov. was collected in two neighbouring sites located on the northern shore of the Sea of Okhotsk, both in the vicinity of Magadan. It seems to inhabit a narrow belt of mudflat maritime marshes, i.e., a monodominant plant association Puccinellietum phryganodis, where it achieves very high abundance levels and is the most common collembolan species. Its occurrence in all other types of marsh in the study area was sporadic. Stach, 1964 Figs 24-29 Syn.: Hypogastrura sheyangensis Jiang, Tang & Chen, 2007. Material. Russia • 3 ♂♂, 9 ♀♀ (slides) and ~ 50 specimens (alcohol); Kuril Islands, Kunashir; 43°42.91'N, 145°33.20'E; wrack beds; August 2017; K. Makarov leg. Several specimens from this material were sequenced (Table 1). Their partial COI genes were amplified and deposited in the GenBank under the sample ID: KY066784-KY066786.

Hypogastrura yosii
Taxonomic remarks. Stach's (1964) original description of H. yosii was based on two specimens collected in forest sites in eastern China. Its real position within the genus had remained unclear until it was recently redescribed from the types (Jia et al. 2011) and synonymised with H. sheyangensis Jiang, Tang & Chen, 2007. The latter species was known from the coastal wetlands of the same region. Specimens from Kunashir Island fit rather well with the existing descriptions, but are significantly smaller, 0.7-0.9 mm vs. "up to 1.5 mm" in Chinese specimens. Both recent descriptions (Jiang et al. 2007;Jia et al. 2011) considered H. yosii as being most similar to two Nearctic species, viz. H. matura (Folsom, 1916) and H. utahensis (Wray, 1953) (now a junior synonym of H. promatro (Wray, 1950) [see Bernard 2015]). This opinion was mainly based on the absence of seta p4 on Abd.4. In fact, this seta, set anteriorly to the p-row (almost aligned with m-setae), may occasionally be absent on one or both sides in many species of the manubrialis-group. This is also rather frequent in specimens of H. yosii from Kunashir Island (Fig. 24). In our view, H. yosii is closely related to the widespread H. manubrialis (Tullberg, 1869). Apart from the absence of a seta m2 on Th.2 (one of the most notable peculiarities of H. manubrialis), the similar number of sensilla on Ant.4, the relatively large PAO with secondary projections at the base of the lobes, and the short anal spines, both species are characterised by an almost identical structure of the maxillary head (except for lam.6, which in H. manubrialis has no denticles in the central part) (Fig. 25) and the absence of a a1 guard on the labial palp (Fig. 26). This short guard is present in most congeners studied, but not in the manubrialis group, in which it is present in only some species: H. arctandria Fjellberg, 1988, H. assimilis (Krausbauer, 1898, H. vernalis (Carl, 1901), and H. promatro, but it is absent from others: H. manubrialis, H. yosii, H. serrata (Ågren, 1904), and H. rangkuli Martynova in Martynova and Chelnokov 1975. Meaningful differences between H. yosii and H. manubrialis appear to be limited by maxillary lam.6 (see above), the presence of seta p' on Ant.1 in H. yosii (a variable character in specimens from Kunashir, where only six of ten specimens examined show p' seta on at least one antenna) and a characteristic mucro shape, i.e., a long and slender mucro without clear lateral lamellae in H. manubrialis or a mucro with an "upturned apex and [clear] outer lamella" in H. yosii (Figs 27-29). Description. Holotype length 0.67 mm. Intravital colour unknown, holotype stored in alcohol for more than 30 years completely lacks dark pigment even on the eye fields. Tegument granulations rather fine and uniform. Ant.4 with a simple apical vesicle and four blunt sensilla (one dorsal and three lateral), rather short and subequal in size, both subapical microsensillum and organite invisible due to the poor condition of the slide. AO typical of the genus, outer sensilla rather short. Ant.1 and Ant.2 with seven and eleven setae, respectively.
Etymology. The name of the new species is derived from the Latin aculeata, meaning spiny, to reflect the shape of the dorsal macrosetae characteristic of X. aculeata sp. nov.
Affinities. The chaetotaxic pattern of X. aculeata sp. nov. allows us to include it into group VI (b), although a forward position of both p1 and p2 on Th. 2 and Th.3 appears to be unique to the new species. There is only one other species in the group that has neither a furca nor a retinaculum, viz. X. osetica Stebaeva & Potapov, 1994. Yet the latter species can easily be distinguished from X. aculeata sp. nov. by the presence of three sublobal setae on the maxillary outer lobe (vs. one in X. aculeata sp. nov.) and a complete chaetotaxy of Th.2 and Th.3 with setae al1 and m3 being present (absent in X. aculeata sp. nov.). Thus latter character (the absence of setae) seems to point to the relations between X. aculeata sp. nov. and several other congeners of the same group that share a functional furca with one (X. betulae Fjellberg, 1985) or two dental setae (X. corticalis Börner, 1901, X. grisea Axelson, 1900, X. hexagona Fjellberg, 1992, and X. laurisilvae Fjellberg, 1992. Among these species, only X. betulae has hitherto been known from the region under study. It appears to be rather similar to the new species in having an identical dorsal chaetotaxy including such fine characters as short sensilla in p-row on Th.2-Abd.1 and the usual absence of setae p5 from Abd.1 and of setae p7 from Abd.4. Ventral chaetotaxy (the absence of p2 from Abd.2 and of some m-setae from Abd.4, the presence of axial unpaired setae on Abd.3), the rather strong anal spines, and the long stout setae on the abdominal tip are also similar. Apart from the complete absence of a furca, X. aculeata sp. nov. differs from X. betulae in having much finer tegument granulation, coarser and more clearly differentiated dorsal setae, only one sublobal seta on the maxillary outer lobe (vs. two sublobals in X. betulae), p1-setae in an anterior position on Th.2 and Th.3, and numerous clavate setae on tibiotarsi (vs. 2-2-2 in X. betulae). The latter character appears to be unique in the genus, but needs verification based on fresh material.
Distribution and ecology. The only known specimen of X. aculeata sp. nov. was found in a typical swampy association of the region, but a search in similar communities in the same area failed to reveal additional material. Taking this into account, as well as some morphological traits of the new species, namely the complete reduction of a furca and the presence of numerous clavate tibiotarsal setae, X. aculeata sp. nov. can be assumed to rather represent a corticicolous, not hygrophilous, species. Paratypes Russia, North-East • 5 ♀♀ and 5 ♂♂, same data as for holotype. The types are deposited in MSPU.
Additional material. more than 500 specimens (alcohol), mainly from the holotype locality; Several specimens from this material were sequenced (Table 1). Their partial COI genes were amplified and deposited in the GenBank under the sample ID: KY066787-KY066789; 9 specimens; same region, AF-51/79: "Geartner Bay, steep slope down to beach with Artemisia, Sedum, Saxifraga, Potentilla, grasses, 20.07.1979";A. Fjellberg leg. Diagnosis. A species from chaetotaxic group II of the genus Xenylla with a general chaetotaxic code of h 1 rt, characterised by a light brownish colour, non-differentiated dorsal setae, and the absence of a prominent cuticular lobe from the subcoxae of hind legs.
Etymology. The new species is named after the famous Norwegian collembologist, Arne Fjellberg, who discovered it in the Magadan Region more than 40 years ago.
Affinities. Using the most recent key to the Asiatic species of the genus (Jia and Skarżyński 2019), X. arnei sp. nov. keys out to X. humicola (Fabricius, 1790), because their general chaetotaxic codes [h 1 rt] are identical. Moreover, both species are similar in many other important morphological traits, namely the labial palp is with four e-  guards (cf. Fig. 38 and Fig. 48), which is not typical of the genus [see Fjellberg (1999)], the virtually identical maxillae (cf. Fig. 40 and Figs 49, 50), three sublobals on the maxillary outer lobe, a retinaculum with 3+3 teeth, a long dens with two setae and a narrow straight mucro. Nevertheless, several fine morphological differences between X. arnei sp. nov. and X. humicola are traceable. Apart from the different colouration, the nondifferentiated setae l1 and l3 on the head (vs. similar in size, but spine-like setae in X. humicola) and the absence of subcoxal lobes from Th.3 in X. arnei sp. nov. (variable in shape, but always present in X. humicola) seem to be the most sound. In addition, the species status of X. arnei sp. nov. is well confirmed by molecular evidence (Table 4). Xenylla arnei sp. nov., together with the widespread X. humicola and the Japanese X. brevispina Kinoshita, 1916 [h 1 rtsv], represent the only known Holarctic members of the species group II (Stebaeva and Potapov 1994). There is also one more form, X. convexopyga Lee, Park & Park, 2005, from the Korean Peninsula, in which the chaetotaxy is rather similar. The chaetotaxic code given for this species in the original description is as follows: krtsv. This code is likely to be not fully correct. The absence of a m3 seta from Th.2 [k] contradicts to fig. 1A in Lee et al. (2005), where this seta is present, albeit marked as m4 (a situation unknown within the genus). In our view, the species rank of X. convexopyga and its separation from X. brevispina need confirmation.
Molecular data. GenBank currently contains COI sequences for only eleven species of the genus (of 140 species known worldwide). The obtained interspecific divergences (Table 4) between all of them, together with X. arnei sp. nov., range between 18.8 to 34.9% (mean 26.5%). Analogous data for X. arnei sp. nov. are 22.2-29.1% (mean 24.9%), which can be considered as an additional argument in favour of its specific status. Nevertheless, such a scant amount of primary data does not allow for any serious statements to be made, but a trend to the absence of parallelisms in molecular and morphological evolution can be traced quite clearly.
Distribution and ecology. Xenylla arnei sp. nov. was collected in two neighbouring sites located on the northern shore of the Sea of Okhotsk, in the vicinity of Magadan. In this region, it was mainly found in various herbaceous meadows at some distance off the coastal line, where its abundance may be very high. Its occurrence in all other types of coastal plat associations in the study area was rather sporadic.