Revision of the endemic Taiwanese millipede genus Aponedyopus Verhoeff, 1939, with descriptions of two new species (Diplopoda, Polydesmida, Paradoxosomatidae). Advances in the systematica of Diplopoda III

Abstract The millipede genus Aponedyopus is endemic to Taiwan and contains three species. All previously described nominal species are considered to represent one species: Aponedyopus montanus Verhoeff, 1939 (the type species), including Aponedyopus reesi (Wang, 1957) and Aponedyopus maculatus Takakuwa, 1942, syn. n. Two further species are described as new: Aponedyopus similis sp. n. and Aponedyopus latilobatus sp. n. The genus is re-diagnosed, all of its three species are keyed, and their distributions mapped.


Introduction
Th e genus Aponedyopus Verhoeff , 1939 was fi rst proposed to incorporate the single species A. montanus Verhoeff , 1939 said to be from the foot of Mt Fuji in Japan. Th e generic diagnosis was very poor, relying on a highly superfi cial resemblance in the gonopod conformation of Aponedyopus to the genus Nedyopus Attems, 1914 (Verhoeff 1939). Takakuwa (1942) was the fi rst to add another species to this genus, A. maculatus Takakuwa, 1942from Taiwan, and, later (1954, questioned the provenance of A. montanus from Honshu, suggesting it had also derived from Taiwan. Wang (1957aWang ( , 1957b described two further species from Taiwan which he fi rst placed in Nedyopus: N. reesi Wang, 1957 andN. jeanae Wang, 1957;then (1963a, 1963b he transferred the latter species to Aponedyopus and referred to it as only a subspecies of A. montanus. Finally, in a checklist of the Taiwanese Diplopoda, Wang (1964) listed the following Aponedyopus: A. montanus montanus, A. m. jeanae, A. reesi, and A. maculatus, the former subspecies quoted as stemming from Japan and being common there, as opposed to the latter three taxa which were said to be endemic to Taiwan. Th us, Wang ignored the previous doubts expressed by Takakuwa (1954) concerning the origin of A. montanus in Japan. He also neglected Miyosi (1959) who had formally synonymized A. jeanae with A. montanus and who had also agreed with Takakuwa that A. montanus had to stem from Taiwan, not from Japan. Hoff man (1980) accepted that the genus occurred in both Taiwan and Japan. Murakami (1993) included A. montanus in the most recent checklist of the Japanese millipedes, but Shinohara and Tanabe (1999) emphasized that the original, type locality of A. montanus, i.e. Mt Fuji, might have been mislabeled. Yet, the genus Aponedyopus still remains on the generic list of Japanese Myriapoda (Tanabe 2001). Jeekel (1968) was the fi rst to properly, however succinctly, re-diagnose Aponedyopus, emphasizing it had nothing to do with the stem-name Nedyopus, because these genera show vastly diff erent courses of their seminal grooves and several other important details of gonopod structure, and belong in diff erent tribes. Yet Jeekel mistakenly listed A. jeanae as a valid species and erroneously believed he was the fi rst to transfer both A. jeanae and A. reesi to Aponedyopus. In fact, Miyosi (1959) had done it before, followed also by Wang (1964). Korsós (2004), in the latest catalogue of the Diplopoda of Taiwan, listed only two species in Aponedyopus: A. montanus and A. maculatus. Concerning the former species, he listed two junior synonyms claimed as new: A. reesi and A. montanus jeanae. However, he must have overlooked Miyosi (1959), who had already synonymized the latter taxon under A. montanus. He also erred in stating that A. maculatus had been described from Ikao, Japan, whereas it had actually been described from Piyanan (= Sihyuanyakou ( ), Datong Township( ), Yilan County ( )), Taiwan (Takakuwa 1942). In addition, not only all of the previous records of these species in Taiwan were summarized, but he also provided some new localities for A. maculatus.
Th e present study reviews the millipede genus Aponedyopus, based on abundant fresh material, including some near-topotypes of one of included species, covering various parts of Taiwan. Th us, the previously described species could be re-assessed, two new species added, and a new synonym established.

Material and methods
New extensive collections of millipedes covering most parts of Taiwan were made between 1989 and 2009, using hand-sorting of the soil and litter. Specimens were preserved in 70% ethanol. External structures were examined and the drawings prepared with a LEICA MZ 16 stereomicroscope, as well as with a HITACHI S2400 scanning electron microscope. Coloration of the specimens is described from alcohol material. Th is material has been shared between the collections of Department of Life Science, National Chung Hsing University (NCHUL), Taiwan; Department of Biological Sciences, National Sun Yat-Sen University (NSYSUB), Taiwan; Department of Life Sciences, National Taiwan Normal University (NTNUL), Taiwan; National Museum of Natural Science (NMNS), Taiwan; Taiwan Forestry Research Institute (TFRI), Taiwan; and Zoological Museum of the State University of Moscow (ZMUM), Russia.

Systematic Account
Genus Aponedyopus Verhoeff , 1939Aponedyopus Verhoeff , 1939Takakuwa, 1954: 49;Jeekel, 1968: 75;Hoff man, 1980: 170;Shinohara and Tanabe, 1999: 681. Diagnosis. Medium-to large-sized Paradoxosomatidae (15-55 mm long, 2.0-5.0 mm wide) with 20 segments. Pore formula normal. Paraterga poorly developed, evident only on segment 2. An evident sternal lobe between ♂ coxae 4; ♂ segment 7 with or without a pair of prominent sternal cones (= spiracles) fl anking gonopod aperture. ♂ tarsal brushes present. Gonopod coxae long, subcylindrical, setose distodorsally, cannula as usual. Telopodites rather long, their distal parts crossing medially in situ. Femorite long, moderately to evidently broadened parabasally on dorsal side, apically separated from postfemoral region by a clear oblique sulcus on lateral side; postfemoral part enlarged at base, tapering thereafter, demarcated from solenophore by a sulcus on mesal side; solenophore shorter than to as long as femorite, curved fi rst ventrad and then dorsad on mesal face, distally holding subparallel to broadened part of femorite; base of solenophore with a small to obvious, apically deeply bifi d lobe; seminal groove fi rst running fully on mesal face of femorite, then turning dorsad near postfemoral part and continuing onto solenomere at base of solenophore on dorsal face; solenomere fl agelliform, long, at most Diagnosis: Diff ers from the other Aponedyopus species in often containing specimens considerably more than 40 mm long, in the considerably longer ♂ legs (usually about twice as long as midbody height), a dentiform process b at the base of the gonopod prefemoral part and, above all, the slender terminal branches (x and y) of the solenophore (Figs 40,42 & 43).
Coloration in alcohol entirely light yellow to dark brown (Figs 1-14). Antennae light yellow to dark brown, increasingly blackish distally, but tip pallid; head to anterior half of epiproct (epi) (Fig. 19), pleurosternal region (ple) (Fig. 18) light yellow to dark brown, prozona (pro) always darker than metazona (meta) (Fig. 18), anterior and hind edges of metazona evidently to slightly lighter brown; posterior half of epiproct, sterna and legs light yellow to orange-brown in ♂.
Distribution: Type material has not been revised, presumably in the collection of the Zoologische Staatssammlung in Munich, Germany.
Th is species is highly variable in size and coloration, and is the most widespread amongst Aponedyopus species in Taiwan. Its distribution covers much of the island and vertically ranges from 175 to over 2,720 m a.s.l. (Map). Name: To emphasize the close resemblance to the next new species. Diagnosis: Being apparently the most similar to A. latilobatus sp. n., based both on several peripheral characters (shorter legs, mostly a smaller body size etc.) and gonopod conformation, it is distinguished by the gonopod lobe b being membranous and lobiform, the terminal branches of the solenophore diff ering in length and crossing each other, with branch x carrying an inconspicuous lobe (see also Key below).
General coloration in alcohol brown to dark brown (Figs 24-27), with a clear pattern of a lighter brown to yellow brown axial stripe consisting of narrower subtriangu-lar spots on proterga and twice as wide central spots on metaterga, these spots growing slightly infuscate, to blackish both towards stricture and posterior half of metaterga; prozona slightly darker than metazona, thus providing a vague cingulate pattern as well; paraterga, legs and venter slightly lighter than background, light grey-brown; head marbled brown, especially well so in vertigial region, genae contrastingly yellowish, a square median spot above antennal sockets contrastingly dark brown; antennae increasingly infuscate, up to blackish distad, distinctly darker at margins, marbled and lighter centrally, only tip contrastingly pallid; both collum and segment 2 with a very faint, yellow-brown, axial line; epiproct uniformly light brown, only very slightly infuscate near base.
Sterna sparsely setose; lamina between coxae 4 setose and emarginate (Fig. 31); segment 7 with a pair of prominent sternal cones (= spiracles) fl anking gonopod aperture; each cross-impression with a transverse sulcus, but without axial groove. Legs (Fig. 27) moderately long and slender, legs 1 to anterior legs of segment 17 with tarsal brushes, thereafter legs broken off in both available ♂♂, each midbody leg ca 1.2 times as long as body height, coxa 2 with a small apical process supporting a gonopore.
Gonopods (Figs 44,45,54 & 55) with process b at base of postfemoral part lobeshaped, membranous, not like a distinct process; l at base of solenophore rather vague; distal part of gonopod deeply bifi d, divided into a longer solenomere (sl), more complex at end and bearing a low terminal lobe, and a slightly shorter, simple, nearly pointed solenophore branch (sph); ends of both branches crossing.

Aponedyopus latilobatus
Diagnosis: Apparently being the most similar to A. similis sp. n., it diff ers in the texture of the tegument (mostly rugulose in A. latilobatus sp. n.) and, especially, in certain details of gonopod structure: lobe l is neither so wide nor membranous, the terminal branches are subequal in length, and the solenomere is supplied with a far more evident terminal lobe (see also Key below).
Coloration in alcohol entirely light brown to brown (Figs 32-35); antennae light brown, growing increasingly blackish distally, but tip pallid; pattern much clearer in ♀, much like in A. similis sp. n.: a light brown, wide, axial stripe from anterior edge of collum to end of epiproct; paraterga and sternites contrastingly lighter brown; legs pallid to yellow; axial line wanting.
Gonopod 56 & 57) with b more like in A. montanus, but l especially indistinct, and distal part of solenophore (sph), albeit also deeply bifi d, having both terminal branches of subequal length, as well as a far more evident terminal lobe on solenomere (sl) not crossing a simple end of sph.
Distribution: Th is seems to be a very local high-montane species in central Taiwan (Map).

Key to Aponedyopus species (based on adult males):
1 Midbody legs about twice as long as body height.

Discussion
Aponedyopus seems to be a small genus endemic to Taiwan. Based on available information, among its three constituent species two are pretty local in distribution, A. latilobatus sp. n. and A. similis sp. n., each restricted to the northern or central, mostly montane parts of the island, respectively (Map). In contrast, A. montanus appears to be extremely widespread, living at various elevations in all parts of Taiwan, including the small islet of Lanyu off the southeastern coast of Taiwan. Whether this species could have been introduced to, and originally described from, Japan, remains open to question. Th ere is only a single example of a basically Taiwanese paradoxosomatid to have become successfully established at least in southern Japan: Chamberlinius hualienensis Wang, 1956 in Kyushu Island and the Ryukyus Kishimoto 1986, 1989;Yamaguchi et al. 2000;Nijima and Arimura 2002).
Th e distribution of Aponedyopus species in Taiwan shows allopatry. Syntopic occurrences are nearly missing, a feature already reported, e.g., for the paradoxosomatid genus Anoplodesmus (Chen et al. 2010), but contrasting with several other adequately known diplopod groups in Taiwan, in which 2-3 congeners are often capable of sharing the same habitat (Chen et al. 2006;Golovatch et al. 2010;Mikhaljova et al. 2010).
Concerning the tribal position of Aponedyopus, it has long been placed in the tribe Tonkinosomatini (Jeekel 1968;Hoff man 1980). However, we rather think that the few basically East to Southeast Asian paradoxosomatid genera forming the tribe Chamberlinini are actually the closest to Aponedyopus. Yet, no formal transfer is advanced here prior to a revision of the type genus Chamberlinius Wang, 1956 (Chen et al., in preparation).