Earthworms newly from Mongolia (Oligochaeta, Lumbricidae, Eisenia)

Abstract Two new megadrile earthworms from the steppes, the first species wholly from Outer Mongolia, are ascribed to the partially parthenogenetic Eisenia nordenskioldi (Eisen, 1879) species-complex. Taxonomic justification of sympatric Eisenia nordenskioldi mongol and Eisenia nordenskioldi onon ssp. n. are supported by mtDNA COI barcodes. The unreliability of molecular differentiation based on voucher names compared to definitive types is again demonstrated, as pertains to the ultimate Eisenia andrei Bouché, 1972 synonym of the Eisenia fetida (Savigny, 1826) sibling species-complex composed of more than a dozen prior names. Similar species described from Northeast China [formerly Manchuria] and North Korea are briefly considered, albeit they are intermittently held in synonymy of cosmopolitan Aporrectodea rosea (Savigny, 1826) along with many other taxa including some exotic lumbricids initially found in India. Japanese and North American lumbricids are also mentioned. Distributions are discussed and an annotated checklist of all nine Siberian/sub-arctic Eisenia nordenskioldi ssp. is appended.

Remarks: The Jeju specimens lack the supposedly characteristic broad striped appearance while specimen S5 is brick red (mtDNA barcodes show 99% agreement). It seems remarkable that S3-4 would agree as they lack pigment. Other specimens with much darker, almost black, segments and contrasting paler intersegment also comply superficially. The whole species-complex requires evaluation with consideration of ICZN compliance as noted in the Discussion.
Etymology. Nominative noun in apposition after sample region where Dadal and the upper Onon River are supposed birthplace and likely final resting place of Temüjin (otherwise known as Genghis Khan).
Remarks. The current taxon differs from previously described subspecies (Tab. 1 and Appendix 2) on its clitellum, TP and tumescences; moreover it appears fertile. Fresher and better preserved material should confirm this analysis. In the meantime, although physically closest to E. n. mongol, it is clearly separated objectively on mtDNA data (Appendix 1). This compares to its sibling species-complex: European Eisenia fetida (Savigny, 1826) vs. E. andrei Bouché, 1972 that is claimed to differ molecularly on enzyme gel electrophoresis, e.g. by Jaenike (1982) based on material from New York, but never yet on respective types of either taxon (see Appendix 1 and Discussion).

Discussion
Interest in natural and acquired species ranges intensifies with global climate concerns. Specific responses to extreme physico-chemical factors are also of interest. Lee (1985 p. 44) reports Ghilarov's claim that Eisenia nordenskioldi revives after long periods of being frozen, with freeze tolerance down to -30°C recorded for E. nordenskioldi (subspecies?) by Holmstrup and Petersen (1997) and Berman and Leirikh (1985). Berman et al. (2002) further report on adaptation to arid conditions. Its sibling species, Eisenia fetida, common at altitude in the Himalayas (Stephenson 1925), may be found in Spitzbergen or Siberia wandering on or under snow (some reports possibly misidentifications of E. nordenskioldi?); and it is also found in deserts (e.g. of Arizona by Gates 1967) and Csuzdi and Pavlicek (2005a, b) recently report it from Mar Saba and Samaria, Israel and Jordan. Eisenia fetida was further located at hot springs on subarctic Iceland and a fumarole at subtropical Raoul Island, N.Z.; its experimental temperature range is given as -2° to +40°C (Lee 1985 tab. 2).
Regarding natural distributions of lumbricid earthworms and species identities, after synonymy of Helodrilus (Bimastus) indicus Michaelsen, 1907, Gates (1958p. 6, 1972) delineated the natural southern boundary of Lumbricidae in Asia north of the Hindu Kush and Karakorum ranges and from Baluchistan west to the Pacific. He thought endemicity of any lumbricid south of Tian Shan and Altai Mts (where giant Eisenia magnifica occurs) into Mongolia or Northeast China would be quite unexpected. Gates (1972 p. 108) said that his synonymy (in Aporrectodea rosea) was not accepted by all authors, indeed Easton (1983 p. 478) resurrected Michaelsen's taxon as Dendrobaena indica, and whereas transfer was questioned by others (cf. genus Healyella), Dr Cs. Csuzdi (2003 pers. comm.) informed that "I have seen the two type specimens. It seems a distinct species with unknown origin". Regardless of its generic status, De. indica or He. indicus can no longer be thought to have been endemic to India, and neither is athecal Kashmiri Al. prashadi (Stephenson, 1922) as noted below.
Although Perel' (1969 p. 62) thought it likely that Allolobophora harbinensis Kobayashi, 1940 belonged in synonymy of Eisenia nordenskioldi, the characters Kobayashi (1940) provided showed similarity to his other three new species that were comparable to Helodrilus (Allolobophora) prashadi Stephenson, 1922: 440, another non-native from India and, after Gates (1958), usually placed in synonymy of Ap. rosea. Kobayashi's data are given in Table 2, albeit all five taxa are currently held in the extensive (four page!) synonym list of Aporrectodea rosea (Savigny, 1826) (e.g., by Gates 1974, Easton 1983, Blakemore 2008a, 2012cf. Tab. 2). Quoting the generic definition by Michaelsen (1900 p. 471), Kobayashi (1940) presumably attributed his taxa to Eisenia as then defined only when the spermathecae were present and in or near the mediandorsal line, otherwise he put them into Allolobophora (including parthenogens?).
Possibly Al. harbinensis is a sexual morph (and therefore an invalid synonym) of Al. hataii. Alternatively, it may represent the amphimixic form of a separate taxon or, equally possible, they are subspecies of either of E. fetida (most likely) or E. nordenskioldi but with spermathecal pores more lateral in cd lines. Nothing of substance separates Kobayashi's Al. jeholensis from his page prior Al. dairenensis so it, at least, should be subsumed. Both have the flared clitella in 29-31 characteristic of Ap. rosea and neither are superficially distinguishable from Aporrectodea rosea itself defined with clitellum in 25,26-31,32 and TP 29-½31,31 or thereabouts, plus several combinations of setal tumescences. Internally Ap. rosea has spermathecae absent or in 9/10/11 dorsally; calciferous glands in 10; U-shaped nephridial bladders and it has a compound typhlosole -see Blakemore ( , 2012. Thus possibly some or all of Kobayashi's taxa, as well as athecal Al. prashadi, may either be Northeast China candidates for Ap. rosea or for parts of the E. fetida and E. nordenskioldi spp.-complexes. Interestingly, Kobayashi (1940 pp. 282-287) describes Eisenia nordenskioldi sub-species as well as both E. rosea and E. fetida from Northeast China! But, since he omits crucial morphological information ("?s" in Tab. 2), more work is therefore required for resolution of all Kobayashi's taxa -extending to DNA analysis of primary types, if locatable and their DNA viable. More probably (topotypic) neotypes will be required -as per Blakemore et al. (2010) -to permit objective comparison with complete and correct identifications on GenBank notwithstanding. Such tasks far exceed the brief of the present study.
For Eisenia nordenskioldi spp-complex, Perel' (1969) separated her E. n. polypapillata from the nominal type and a similarly unpigmented, E. n. pallida (Malevic, 1956) on the basis of its numerous papillae between the male pores and clitellum, and on the wider distance separating the spermathecal pores from the mid-dorsal line (Tab. 1). Dr Perel (pers. comm. Dec. 2012) now suspects both subspecies are variations of the same taxon, however this too would require reference to the earlier pallida and acystis types (if locatable).
As with E. fetida, mere colour differentiation is probably inadequate. Kobayashi (1940), whose taxa were subsequently combined irrespective of their pigmentation, said typical E. nordenskioldi somewhat resembled E. fetida but were not quite so banded intersegmentally. In contrast, Zicsi (1972) noted his E. nordenskioldi specimens reddish in life, when preserved were colourless. Thus wide intraspecific colour variations seem permissible in parts of E. nordenskioldi too.
Perel' (1987) states: "..Eisenia nordenskioldi in southern Kazakhstan, Siberia and the Far East is represented by the poorly pigmented form pallida. The typical form is significantly more widespread, in Siberia reaching the regions of the far north and also occurring in the eastern and south-eastern parts of the European USSR" and Perel' (1997 p. 22) gives the location of dubious Eisenia nordenskioldi pallida morph or subspecies in "..Китая и на севере Корея." (= China and in northern Korea). These citations by Perel (1979Perel ( pp. 75, 267, 1997 may be mistaken if priority yields to Eisenia acystis , thus leaving Eisenia nordenskioldi pallida Malevics, 1956 as species inquirendum. Historical reports of the nominal taxon from the Azores and Hawaii are probable misidentifications with E. fetida (as noted by Michaelsen, 1900 p. 476); while Garman (1888 p. 73) said that E. nordenskioldi was: "Obtained by Eisen in Siberia; credited to North America by Vejdovsky", i.e., its USA credit was mistaken too.
Confusion between these sibling species may have been common. Both Michaelsen ( , 1910 and Gates (1972 p. 103) recognized variability of (parthenogenetic and/or polyploidal) morphs and close relationships of Siberian Eisenia nordenskioldi (Eisen, 1879) with European E. fetida (Savigny, 1826), Gates saying they were "indistinguishable specifically from each other by any of the characters of the classical system" and differing substantially only in the number of atyphlosolate posterior segments. The whole Eisenia fetida species-complex yet requires evaluation with consideration of ICZN compliance. For example, as noted above, Jaenike (1982) avoided types and overlooked the synonyms with priority over E. andrei, the first being E. semifasciatus (Burmeister, 1835) which has not yet been tested and neither have any of Kobayashi's species as noted herein. Moreover, at least Stop-Bowitz (1969, tab. V) maintains Scandinavian E. fasciata Backlund, 1948 which is often included in E. fetida synonymy by most authors along with ca. 14 other names, but more often than not (especially in chemical/molecular studies by non-taxonomists) these available synonyms are completely overlooked (see also the discussion in Blakemore et al. 2010). This notion, that components of the E. fetida and E. nordenskioldi spp. complexes are indistinct, is gradually being falsified by refined genetic information complementing the morphology of taxa under rules of ICZN (1999) that disallows nomenclatural availability to varietal forms, morphs or races. However, further considerations are, firstly, that genetics only reveals a part of the information on a taxon while a morphological character is often controlled and manifest from interplay of several genes throughout the organism's ontogeny and phylogeny (with ontogeny defined as the history of structural change in any biotic entity whether a cell, an organism, or a population of organisms, i.e., a species). Secondly, regardless of data being based on DNA or morphology, or on both of these, it is only the condition pertaining to the ICZN (1999) defined type-specimen that defines the scientifically-named species. Hence a chronic confusion of all E. fetida/andrei results -see Blakemore (2006Blakemore ( , 2012bBlakemore ( , 2013a and Blakemore et al. (2010), the latter while also providing a model from the first COI barcode of an earthworm's neotype, comments on the shortcomings of all previous molecular studies. Just as Blakemore (2011) observed regarding a New Zealand paper: "as with several previous molecular phylogenetic works, the only errors in their otherwise informative study are the names".