Subdivision of the tribe Oligaphorurini in the light of new and lesser known species from North-East Russia (Collembola, Onychiuridae, Onychiurinae)

Abstract The paper is devoted to a taxonomic review of Oligaphorurini from the north-eastern part of Palaearctic. Two new species, i.e. Oligaphorura ambigua sp. n. and Oligaphorura duocellata sp. n., are described. Four species, Oligaphorura nataliae (Fjellberg, 1987), Oligaphorura interrupta (Fjellberg, 1987), Oligaphorura pingicola (Fjellberg, 1987), and Micraphorura alnus (Fjellberg, 1987), are redescribed on base of the types and new material, and remarks on other species known for the region, Oligaphorura groenlandica (Tullberg, 1876), Oligaphorura ursi (Fjellberg, 1984), Oligaphorura aborigensis (Fjellberg, 1987), and Micraphorura absoloni (Börner, 1901), are given to clarify their generic affiliation. Finally, merits and disadvantages of the current subdivision of the tribe are discussed and a key to the northern species of the tribe is provided.


Introduction
Two undescribed species of the tribe Oligaphorurini from the upper reaches of Kolyma River (North-East Russia, Magadan region) do not fit the current generic subdivision of the tribe, which is mainly based on revisions made by Pomorski (1996) and Weiner (1996). Other east Palaearctic species, described by one of the authors (A. Fjellberg) from the same region, also need a critical review to fix their generic affiliation. Below, the new species are described, others are redescribed in more detail, and finally the current generic subdivision of the tribe is critically analyzed. Bagnall (1949) was the first author to recognize Oligaphorurinae (as a subfamily of Onychiuridae). He split it into four genera: Archaphorura, Micraphorura, Oligaphorura and Dimorphaphorura based on five species only, the sixth species, described in the same paper, is now considered as a synonym. Since then many other species have been established, often in shifting generic associations. At present, according to the database of Collembola of the World (Bellinger et al. 1996(Bellinger et al. -2014 approximately 50 species of the tribe are known. Using different approaches Weiner (1996) and Pomorski (1996) retained the four original genera described by Bagnall. One new genus, Chribellphorura, was established by Weiner (1996) for Onychiurus allanae Christiansen & Bellinger, 1980, displaying a unique set of characters. However, this generic framework bears internal contradictions and does not cope with the known morphodiversity of the species. Both Pomorski and Weiner (op. cit.) based their diagnoses on the gradual reduction of the furcal field on the sternum of the fourth abdominal segment. Pomorski examined the first instar juveniles, while Weiner used adults. Both authors studied a rather limited set of mainly European species. Table 1 summarizes the diagnostic characters separating the genera and which species were involved.

The current subdivision of Oligaphorurini
Recently Shvejonkova and Potapov (2011) described three new species of Oligaphorurini which did not possess anal spines, a feature which was characteristic only to the genus Archaphorura. Nevertheless the species were assigned to the genera Oligaphorura and Micraphorura, an action which brings the above diagnostic scheme to a state of collapse. However, the cited authors did not establish the synonyms which would have been a natural consequence.
In 2014 a complete revision of the genus Dimorphaphorura has been undertaken (Weiner and Kaprus' 2014). The authors of this revision described six new Palaearctic species of the genus, redescribed and clarified generic affiliation of a number of other known species, and defined diagnostic characteristics of the revised genus. According to the diagnosis provided, it differs from other genera of the tribe in the organization of the furcal area (see Table 1). As a result the majority of Palaearctic species previously treated as Micraphorura have been transferred to Dimorphaphorura. The two new species described below introduce further chaos in the existing generic system.  Weiner (1996) Apical vesicle on Ant.4  Abbreviations A, AB, AC and ABC -four types of labium in Onychiuridae in accordance with the presence of thickened and blunt-tipped setae on corresponding labial papillae (Fjellberg 1999) ABD -the fifth type of labium in Onychiuridae (Shvejonkova and Potapov 2011) Abd.1-6 -abdominal segments A-B, T-setae, setae M and Y -tibiotarsal setae (Deharveng 1983) Ant.1-4 -antennal subsegments AO -antennal organ on Ant.3 a 0 , m 0 , and p 0 -unpaired axial setae on terga CNC -Canadian National Collection (Ottawa) d 0 -unpaired axial seta on area frontalis of the head ma-, mm-and mp-row -anterior, medial and posterior rows of setae on manubrial field (Weiner 1996) ms -microsensillum MSPU -Moscow State Pedagogical University PAO -postantennal organ pso -pseudocellus(i) psx -parapseudocellus(i) q-setae -proximal setae on furcal field of Onychiuridae (Pomorski 1996)  Description. Colour white. Size of adults 0.73-0.92 mm. Body slender and elongated, Abd.3-4 clearly widened, Abd.6 short and hardly visible in dorsal view (Fig. 1), anal spines not developed (Fig. 3). Antennae about as long as head, Ant.4 not wider than Ant.3 (Fig. 5). Ant.4 with spherical subapical organite surrounded by cuticular papillae (Fig. 6), basal microsensillum present on level with proximal whorl of setae (Fig. 5). AO consisting of 4 finger-like papillae, 2 sensory rods, 2 smooth sensory clubs clearly differing in shape ( Fig. 7), 5 guard setae and a lateral microsensillum (Fig. 5). Ant.1 and 2 with (8)9 and 14-15 setae respectively. PAO smaller than nearest pso, usually with 3 subequal lobes (Fig. 8). Labrum with 7 setae and 4 prelabral ones. Apical part of labium with thick terminal setae on papillae A, B and C (ABC-type), 7 long and 4 spiniform guard setae, and 6 proximal setae (Fig. 4). Basal fields of labium (mentum and submentum) with 4 and 5 setae, hypostomal complex of usual shape. Maxillary palp simple, with 2 sublobal setae.
Affinities. This new species resembles two congeners recently described from the European part of Russia, namely Oligaphorura humicola Shvejonkova &Potapov, 2011 andO. kremenitsai Shvejonkova &Potapov, 2011. All three species lack anal spines and have no cuticular fold on the sternum of Abd.4. Apart from this, the former is characterized by a set of pso on both dorsal and ventral sides on a body, as well as on upper subcoxae identical to that in O. ambigua sp. n., and also has no ms on Th.3 and ventral psx. O. kremenitsai differs from both O. ambigua and O. humicola having more pso on Th.2-3 (42/144/33354 as a whole). Some differences like the uncommon position of anterior pso on head and submedial ones on Abd.4-5 in O. humicola and O. kremenitsai or their reduced tibiotarsal chaetotaxy may be a result of small body size (< 0.6 mm). Nevertheless, the palp structure (ABD-type) in humicola/kremenitsai and the loss of labial papilla C probably reflects a certain genetic distance.
The presence of only four papillae in AO is also an unusual condition in the tribe, shared only with Micraphorura absoloni (Börner, 1901), Oligaphorura palissai (Yosii, 1971) and Dimorphaphorura sophyae Weiner & Kaprus', 2014. The clear cuticular papillae on antennal tip are also quite characteristic.
Etymology. The name reflects the uncertain generic position of the new species. Distribution. Known only from the type locality, the alpine belt in the upper reaches of Kolyma river.  , 27.vii.1979, leg. A. Fjellberg (MSPU).
Affinities. Several uncommon features, like 2+2 pseudocelli and few setae on Th.1, the presence of pseudocelli on several abdominal sterna, the absence of setae on thoracic sterna, and a furcal remnant in the form of a finely granulated area with 4 small setae behind it, permit easy identification of the new species. In addition to O. duocellata sp. n. nine known species of the tribe possess pseudocelli on several abdominal sterna. Three of them, Archaphorura serratotuberculata (Stach, 1933), A. alavensis Simón & Luciáñez, 1994, and A. marcuzzii (Cassagnau, 1968) are usually considered as representatives of the genus Archaphorura due to the absence of anal spines. The presence of ventral pso in the former species is uncertain as specimens from Moscow vicinity contrary to those from Poland (see Pomorski 1998)  The only known species of the tribe with AS and pso on several abdominal sterna (M. multiperforata (Gruia, 1973), M. uralica (Khanislamova, 1986), are within Micraphorura on the www.collembola.org or treated as Dimorphaphorura (D. olenae Weiner & Kaprus', 2014). M. multiperforata is a unique species with dorsal pso multiplication, whereas M. uralica seems to be the most similar to O. duocellata sp. n. having also more than 2 pso on subcoxae (a unique character) and no setae on thoracic sterna, a character which is known only for species from eastern parts of Asia and North America, i.e. Oligaphorura nuda (Fjellberg, 1987), O. judithae (Weiner, 1994), O. linderae (Weiner, 1994), O. montana Weiner, 1994, O. pseudomontana Sun & Wu, 2012, and O. chankaensis Sun & Wu, 2012 Etymology. Named after the presence of 2+2 pso on Th.1, a character previously unknown for the tribe.
Distribution. Known only from the type locality.

Remarks.
Juveniles from Magadan (NE Russia) have a furcal field with 2+2 setae behind the cuticular furrow, followed by 3+3 q-setae (Fig. 27). This pattern is in a full accordance with what Pomorski (1996) reported from European populations. In adults a few additional setae (usually in asymmetric positions) may appear between the primary rows of the juvenile, obscuring the original pattern (Fig. 28). The formula of the parapseudocelli (psx) in specimens from Magadan is also the same as Pomorski (1998) noticed from Europe: 0/000/1101, absent on subcoxae.
Remarks. Originally described as Onychiurus (Archaphorura) nataliae, the species is now listed under Micraphorura on www.collembola.org. Nevertheless the chaetotaxy of the manubrial field in juveniles of this species is identical with that found in north-ern partenogenetic populations of O. groenlandica (Tullberg, 1876) (cf. Fig. 32 and Fig. 18). Adults usually have a pattern with four setal rows behind the cuticular furrow (Fig. 31), which considered being typical for Oligaphorura.
In the interactive key on www.collembola.org the species keys out with Oligaphorura interrupta (Micraphorura on www.collembola.org) which can easily be distinguished by higher number of abdominal pso, presence of ms on Th.3 and absence of ventral setae on Th.2.
Five known species of the tribe possess the same number of dorsal and ventral pseudocelli as nataliae, i.e. O. pingicola (Fjellberg, 1987), O. koreana (Weiner, 1994), Dimorphaphorura raxensis (Gisin, 1961), D. chatyrdagi (Kaprus', Weiner & Pomorski, 2002), and D. sanjiangensis Sun & Wu, 2012. O. nataliae differs from the above Oligaphorura species (O. pingicola and O. koreana) in having no ms on Th.3. D. raxensis according to Weiner and Kaprus' (2014) has 9 distal setae on tibiotarsi and ABC type of labium. D. chatyrdagi can easily be distinguished due to strongly reduced tibiotarsal chaetotaxy (with 5 distal setae) and the presence of ms on Th.3. D. sanjiangensis, recently described from northern China, can be separated from O. nataliae due to different type of labium (A versus AC in O. nataliae), the presence of psx on abdominal sterna (0/000/122201+1 m ), and identical number of tibiotarsal setae on all legs (20-20-20) which is very characteristic if it is correct. Apart of this, all three species of Dimorphaphorura should have no more than 5+5 setae in the manubrial field whereas even first instars of O. nataliae possess 6+6 setae.
The presence of 7 long guard setae of labium in such small species as O. nataliae is an uncommon character in the tribe and needs additional confirmation being seen in few specimens. Only five other Asiatic species, D. sanjiangensis, Micraphorura changbaiensis Sun & Wu, 2012, O. aborigensis (Fjellberg, 1987) and the two new Oligaphorura species described above, share this character with O. nataliae whereas 18 species of the tribe are known as having only 6 long guards. (Fjellberg, 1987), comb. n.
Remarks. The above redescription completely matches the original one, although Fjellberg (1987) did not mentioned ventral pso on Abd.4. The species was recently redescribed by Weiner and Kaprus' (2014). The only clear difference with this description is the number of labral setae which states as being full (4/3-4-2). The authors treat the species as Dimorphaphorura because their specimens had no secondary setae in the manubrial field (only 5+5 setae in all). In our largest specimens from Magadan the manubrial field has several additional mm-setae (the holotype has 14 setae on the manubrial field in total) which illustrates the weak distinction between Dimorphaphorura and Micraphorura.
Remarks. The number of pseudocelli in the species appears to be more variable than stated in the original description by Fjellberg (1987), even within the region of the type locality. That is why Nearctic Oligaphorura nuda, characterized by increased number of abdominal pso, appears to be hardly separable from O. interrupta despite their different generic positions on the www.collembola.org. Nevertheless the chaetotaxy of manubrial field in O. interrupta is identical to that of Oligaphorura groenlandica  and clearly differs from the pattern typical of Micraphorura absoloni (Figs 26-27). The structure of manubrial field in the nuda holotype (CNC 165047, type No 20103) also indicates its position within the genus Oligaphorura.
The presence of ventral setae on Th.3 in O. interrupta was used by Fjellberg (1987) as an additional diagnostic character to separate O. interrupta and O. nuda (setae absent). However, new material of O. interrupta from various regions of eastern Palaearctic shows this character to be invalid. Some specimens of O. interrupta may also be completely devoid of ventral setae on thorax. More material is evidently needed to clarify the real relationships between these two species.
Only one other known species of the tribe shares the absence of pso on Th.1 combined with presence of 5 pso on Abd.4 with O. interrupta and O. nuda: Oligaphorura reversa (Fjellberg, 1987). This characteristic species differs from the above-mentioned species in having an unusual position of the dorsal pseudocelli on Abd.1-3: the medial pso a is set in a posterior position, clearly behind submedial pso b.

Oligaphorura groenlandica (Tullberg, 1876)
Lipura groenlandica Tullberg, 1876: 41. Oligaphorura groenlandica (Tullberg): www.collembola.org Remarks. Pomorski's (1996) description of the furcal area of the first instar was as follows: …q-chaetotaxy -3 chaetae, area furcalis with 2+2 setulae below cuticular furrow and 2+2 setae at base [all together 4+4 setae]. It was based on a single specimen from a bisexual population from Wolin Island on the Polish shore of the Baltic Sea. Weiner's description (1996): …small, finely granulated cuticular fold or quite a deep pocket with 2 setae on its posterior edge, sometimes with 1+1 additional setae and two other dental setae posteriorly, with manubrial setae on both sides and with other manubrial setae usually in two rows is more complicated. According to the interactive key on the www. collembola.org Oligaphorura should have two dental setae on the fold or posteriorly and three manubrial rows of setae behind them. In fact, the type species of the genus, O. groenlandica (or more correctly the most common parthenogenetic form of this species) has no cuticular fold or clear furrow, just an area with fine granulation in anterior third of the sternum of Abd.4 (Fjellberg 1998). In adults, the position of setae of the furcal area is rather irregular due to weak polychaetosis (Fig. 17) and juveniles clearly differ from that described by Pomorski with only 3+3 setae in front of the 3+3 q-setae (Fig. 18).
Unfortunately, this parthenogenetic form is not the only one present in the northern areas of the Palaearctic. On Taimyr Peninsula and Novosibirsk Islands another bisexual form was found. Probably the same (or similar) form exists in southern Norway (Fjellberg 1998) and Poland (Pomorski 1998). Its furcal area is more similar to the described pattern for O. groenlandica by Pomorski (1996) with cuticular fold and 4+4 setae between the fold and q-setae in the first instar juveniles (Fig. 26). In adults, two "dental" setae set in front of three irregular manubrial rows of setae (Fig. 25). These two forms are very similar and apart from the furcal area, differ only in size (the partenogenetic form is larger) and in differentiation of the medial setae on Abd.5: "microsetae" m 1 (thin and pointed) are much longer than macrosetae a 1 and p 1 (straight and truncate) in the true parthenogenetic O. groenlandica. Bisexual specimens usually have m 1 curved and short and a 1 and p 1 long and straight. There are also some differences in psx formulas: 10/000/222201+1 m , upper subcoxae with 2-2-2 psx in the bisexual form and 10/000/222101 m , subcoxae 1-1-2 in the parthenogenetic one. Unfortunately the number of psx in the latter form is not stable. Some specimens lack postlabial psx or one of psx on anterior abdominal sterna, others may have additional psx on Abd.4 or on paired anal lobes; and anterior psx on subcoxae of fore and middle legs can be just invisible due to position. Pomorski (1998) gave slightly different formula for the Polish specimens: 1/000/122101 m . Thus several similar forms do exist in Palaearctic, but the real O. groenlandica described by Tullberg from Greenland and Svalbard probably belongs to the main parthenogenetic form with circumpolar distributional range lacking cuticular fold on the sternum of Abd.4. (Fjellberg, 1984) Onychiurus ursi Fjellberg, 1984: 71. Oligaphorura ursi (Fjellberg): www.collembola.org

Oligaphorura ursi
Remarks. Contrary to O. groenlandica, O. ursi, another northern circumpolar species of the genus, is common in the Magadan region inhabiting different wet sites above the tree-line. Recently the species was redescribed on the basis of specimens from northern China (Sun and Wu 2012a). We have some doubts about the identity of the Chinese and northern populations. Northern specimens usually have 6 long and four spiniform guard setae on the labial palp [versus 11 in Chinese specimens], ventral psx 10/000/212201+1 m with frequent variations [versus 0/000/122200 in the Chinese ones] and at least one psx on each subcoxae, most usually 1-2-2 [versus completely absent]. Apart from this, it was said that the Chinese specimens had an identical number of setae on all tibiotarsi [versus 20-20-19 setae in northern populations].
Oligaphorura aborigensis (Fjellberg, 1987) Onychiurus (Archaphorura) aborigensis Fjellberg, 1987: 285. Unfortunately the types of the species were partly damaged and no additional specimens were found in the available material from the vicinity of Aborigen field station. So, only few additional details can be added to the original description.
Labium with thick terminal seta only on papilla A, 7 long guard setae and 6 proximal setae, basal fields with 4+6 setae. Tibiotarsi with complete set of setae (20-20-19): each distal whorl (A+T) with 11 setae, whorl B with 7-7-6 setae, setae M and Y present on all tibiotarsi. Furcal fold straight and comparatively small, situated in mid-section of Abd.4, furcal field in the only seen juvenile with 4+4 setae between proximal q-setae and the cuticular fold (as on Fig. 24), adults with some additional setae in intermediate position forming 4 more or less regular rows as typical for other Oligaphorura.
The species is well defined due to the absence of sublobal setae on the maxillary outer lobe (a unique character for the tribe or even for Onychiurinae), strong differentiation of dorsal setae and the pseudocellar formula (32/133/33353) which is not especially common for the tribe being shared with only D. pseudoraxensis (Nosek & Christian, 1983), O. sabulosa Babenko, 2008, andD. jingyueensis Sun &Wu, 2012. All of them have the usual two sublobals on the maxillary palp and the macrosetae being much shorter and finer than in O. aborigensis. Apart from this the two former species are characterized by the absence of ms on Th.3. (Fjellberg, 1987) Onychiurus (Archaphorura) pingicolus Fjellberg, 1987: 285. Oligaphorura pingicola (Fjellberg) Redescription. Colour white. Granulation distinctly enlarged on Abd.6 and on head. Size 1.0-1.1 mm. Body shape cylindrical. Antennae about as long as head, Ant. 3-4 broad, club-like. Ant.4 with a subapical spherical organite and a microsensillum located in proximal row of setae. AO consists of 5 long and thin papillae, two sensory rods, two granulated sensory clubs (internal almost straight, external much larger and bent), 5 guard setae and a lateral microsensillum which set below the organ. Ant.1 and 2 with 8-9 and 15-16 setae respectively. PAO with 2-3 elongated lobes, much longer than nearest pseudocellus. Labrum with 4/5-2-2 setae. Apical part of labium with thick terminal setae on papillae A and C, 6 long guard setae and 6 proximal setae present, basal fields with 4+(5)6 setae. Maxillary palp simple with two sublobal hairs.
Remarks. The above redescription is in full accordance with the original one, adding a few details. While originally described from Alaska, Fjellberg (1987) also remarked that two specimens of the main form are also seen from alpine meadows at Aborigen, USSR (Magadan Reg.). In fact the species seems to be widespread and common not only in the Magadan Region but also in inner parts of the eastern Palaearctic (Suntar-Khayata Mts. Range, Yakutia). Fjellberg (1987) mentioned two distinct forms for Alaska differing in mutual position of setae on Abd.5 and in level of granulation. Only the main form seems to be present in the eastern Palaearctic.
Oligaphorura pingicola shares the number of dorsal and ventral pseudocelli with at least five known species of the tribe, namely O. koreana, O. nataliae, Dimorphaphorura raxensis, D. chatyrdagi, and D. sanjiangensis. O. koreana is very similar to O. pingicola, differing by fewer tibiotarsal setae (19-19-18 versus 20-20-19) and by absence of psx ("indistinct"). The absence of psx is also characteristic for O. nataliae which differs from O. pingicola in having 7 long guard setae on labial palp and absence of ms on Th.3, as well as 2 setae of the proximal row on labrum (4/7 as a whole). Dimorphaphorura raxensis has 9 distal setae on tibiotarsi, full number of labral setae and ABC type of labium (Weiner and Kaprus' 2014). Dimorphaphorura chatyrdagi can easily be distinguished due to reduced tibiotarsal chaetotaxy with only 5 distal setae, whereas D. sanjiangensis apart from the chaetotaxy of the sternum of Abd.4 differs in labium type (A versus AC) and identical number of tibiotarsal setae on all legs (20-20-20 versus 20-20-19).
Oligaphorura tottabetsuensis (Yosii, 1972), a species known from northern Japan, probably also belongs to the same group although the reported number of dorsal pseudocelli is slightly different (32/033/33333). The species is in need of redescription.

Discussion
The morphological characters being widely accepted as separating genera of Onychiuridae involve the shape of the postantennal organ (PAO), structures of the antennal organ (AO), tibiotarsal chaetotaxy, arrangement of the pseudocelli (pso), presence/ absence of anal spines, distribution and shape of sensory setae on the body, and the gradual reduction of the furca. In our view a genus diagnosis based exclusively on reductional stages of the furca is dubious for at least two reasons: (1) similarity in reductional stage may represent a convergence achieved independently from distant phyletic lines, resulting in a polyphyletic or paraphyletic assemblage of species; (2) many collembolan genera (Xenylla, Folsomia, Folsomides, Scutisotoma, etc.) cover species with a wide range of furcal reduction, but are still accepted as natural genera which no one would split. In Collembola at least the initial stages of furcal reduction are clearly of adaptive nature, reflecting a shift from surface activity to life in deeper strata where jumping ability is restricted. Although the species under discussion have a furca which is no longer functional, the adaptive character of the reduction probably masks the underlying genetic relationships. Moreover, the practically identical furcal remnant of the Onychiuridae genera Supraphorura Stach, 1954 andPsyllaphorura Bagnall, 1948 is obviously not a good proof for any close relationship. Bagnall's (1949) original diagnoses of four genera of Oligaphorurini were more species than genus diagnoses. Re-establishment of these genera by Weiner (1996) and Pomorski (1996) was based on other diagnostic characters and involved more species but created some taxonomic problems which are not yet solved. According to these authors the four principal genera may be recognized as follows: Archaphorura and Oligaphorura differ by absence of anal spines in the former, presence in the latter. Both have identical furcal fields, differing from the two other genera by an additional row of setae, even in the first instar juvenile. Dimorphaphorura has the same chaetotaxy of the furcal field as a juvenile Micraphorura, whereas adults of the latter have a few (1-4) additional intermediate setae. In practice juveniles of Archaphorura and Oligaphorura are easily separated, also from juveniles of Dimorphaphorura/Micraphorura, while juveniles of the two latter are inseparable by the furcal field. Sorting out the generic affiliation of adults is much more difficult.
On the www.collembola.org there is an interactive key which proposes the following characters for identification of Oligaphorurini genera.
Chribellphorura: antennal tip with a retractive papilla, tibiotarsi with clavate setae in distal whorl; Archaphorura: Abd.5-6 fused dorsally, Ant.3-4 fused, anal spines absent; Dimorphaphorura: furcal rudiment in a form of finely granulated area; Oligaphorura: furcal rudiment in a form of cuticular furrow or small fold; chaetotaxy: 2 dental setae on the fold or posteriorly and three manubrial rows of setae behind them; Micraphorura: similar to Oligaphorura but without 2 dental setae, so only three rows of setae can be distinguished, mm-row with 4-6 setae.
In summary, Archaphorura has a unique character combination, Dimorphaphorura has no furcal fold or furrow, and Oligaphorura has an additional row of the setae on manubrial field compared with Micraphorura. The monotypic genus Chribellphorura is unique and needs no further discussion to be distinguished.
This adequate but probably too simplified scheme was neglected by Shvejonkova and Potapov (2011) who included three new species without anal spines and furcal fold not in Archaphorura but in Micraphorura (M. stojkoae) and Oligaphorura (O. kremenitsai and O. humicola). As a result Archaphorura lost its main diagnostic feature (absence of anal spines), as did Dimorphaphorura (furcal rudiment in the form of a finely granulated area). The authors considered Archaphorura to be a good genus not due to the absence of AS, but because of the peculiar antennae (fused Ant.3 and 4, subapical position of AO and ms of Ant.4 hidden under long papillae) and the fused Abd.5-6. There is also one neglected diagnostic character state of A. serratotuberculata, the type species of the genus, namely the absence of M-setae on tibiotarsi. This very character was registered in two rather remote European regions, Fennoscandia (Fjellberg 1998) and Moscow vicinity (new data). In all other Oligaphorurini with known tibiotarsal chaetotaxy this seta is present. Unfortunately there is still a number of species in which tibiotarsal chaetotaxy is not completely known and the character may end up as non-diagnostic for Archaphorura. The exact definition of A. serratotuberculata is obscure (Shvejonkova and Potapov 2011) and several species may be involved.
The recent revision of the Palaearctic species of Dimorphaphorura by Weiner and Kaprus' (2014) defined the genus more legibly and added a further criterion distinguishing Dimorphaphorura from other genera of the tribe -the absence of so called "dental" setae. In our view this character is rather subjective as setae set just below cuticular trace of furca differ (in size or sockets) from other sternal setae on Abd.4 only occasionally even in M. absoloni, can hardly represent a reliable criterion. The two new species described in the present paper further complicate the situation as one of them lacks AS but has long, not club-like, antenna and both have "furca" in the form of a "finely granulated area" (typical of Dimorphaphorura) but with 2+2 small setae in two rows below it as in Oligaphorura (Figs 15-16, 29-30).
In fact the diversity of manubrial chaetotaxy patterns in Oligaphorurini seems to be much higher than postulated so far, which obscures the current generic subdivisions. Thus four different patterns were found in juveniles of the northern species of the tribe: apart from 3+3 proximal q-setae the furcal field may have 2+2 setae (alnus, absoloni, Fig. 27), 3+3 (groenlandica,nataliae,interrupta,32) or 4+4 setae (most studied Figs 21,24). In the latter case there are at least two patterns with a different mutual position of setae (cf. Fig. 21 and Fig. 26). The fifth variant with only 1+1 setae is known for the first instar of Dimorphaphorura daii . The number of secondary setae appearing during ontogenesis on the furcal field is surprisingly low, usually 1-3, rarely more. Nevertheless, the position and the number of these secondary setae are not stable within a species. We have probably never seen any adult specimens with completely symmetrical chaetotaxy of furcal field when there are any secondary setae present.
The present generic framework for Oligaphorurini is probably unique -difficult to use and hardly reflecting real relationships. There is great temptation to return to a single genus Archaphorura as Christiansen and Bellinger (1980Bellinger ( , 1998 and Fjellberg (1987) had done. Nevertheless we realize that such a pooling of all species of the tribe within a single unit obviously contradicts recent taxonomic traditions. Despite being unsatisfactory the most realistic alternative is to keep the current "five genera" system, admitting that this solution is clearly provisional and does not fully reflect a natural generic affiliations of the species pool. Hopefully future genetic studies (barcoding) may contribute to solve these problems.