Revision of the Taiwanese millipede genus Chamberlinius Wang, 1956, with descriptions of two new species and a reclassification of the tribe Chamberlinini (Diplopoda, Polydesmida, Paradoxosomatidae, Paradoxosomatinae)

Abstract The millipede genus Chamberlinius is basically confined to Taiwan, with only one of the four known species presumably introduced to southern Japan. Both previously known species are redescribed, based on new material: Chamberlinius hualienensis Wang, 1956 (the type species) and Chamberlinius piceofasciatus (Gressitt, 1941), the latter being a new subjective senior synonym of Chamberlinius shengmui Wang, 1957, syn. n. Two further congeners are described as new: Chamberlinius pessior sp. n. and Chamberlinius sublaevus sp. n. The genus is re-diagnosed, all of its four species are keyed, and their distributions mapped. The tribe Chamberlinini is reclassified and, based on gonopod traits, shown to comprise the following five genera: Chamberlinius Wang, 1956, Haplogonosoma Brölemann, 1916, Riukiupeltis Verhoeff, 1939, Aponedyopus Verhoeff, 1939 and Geniculodesmus Chen, Golovatch and Chang, 2008.


introduction
The genus Chamberlinius Wang, 1956 was first erected, based on the single species C. hualienensis Wang, 1956, taken from Taruko, Hualien, Taiwan (Wang 1956). A year later, Wang (1957) described another new congener, C. shengmui Wang, 1957, and he also proposed a new subfamily, Chamberlininae. Jeekel (1968) placed Chamberlinius in the tribe Sulciferini, subfamily Paradoxosomatinae. He compared the illustrations of Orthomorphella pekuensis (Karsch, 1881), the type species of Orthomorphella Hoffman, 1963, which Hoffman (1963 had validated, with the poor line drawings of both Chamberlinius hualienensis Wang, 1956 andC. shengmui Wang, 1957 as presented by Wang (1956Wang ( , 1957, to conclude that Chamberlinius was a senior synonym of Orthomorphella. This opinion has since been shared by Wang and Mauriès (1996). Furthermore, Jeekel (1968) referred to Chamberlinius as containing four species: C. cristatus (Takakuwa, 1942), C. hualienensis Wang, 1956 (the type species), C. pekuensis (Karsch, 1881) and C. shengmui Wang, 1957. Soon after that, Hoffman (1973), disagreeing with Jeekel (1968), resurrected Orthomorphella as an independent genus, removed Chamberlinius from the tribe Sulciferini, and downgraded the subfamily Chamberlininae to the tribal status, Chamberlinini in Paradoxosomatinae. This tribe was stated to encompass only Chamberlinius and Riukiupeltis Verhoeff, 1939. He was also the first to assign C. piceofasciatus (Gressitt, 1941), another species from Taiwan, to Chamberlinius. He redescribed both C. hualienensis and C. piceofasciatus, simply listing C. shengmui as a third congener. More recently, Hoffman (1980) spoke already of "about five species" in this genus, which is a slightly puzzling estimate because only the above trio has hitherto been known to compose Chamberlinius.
Regrettably, since all of the diplopod material Wang (1956Wang ( , 1957 claimed to have deposited in the Department of Zoology (now incorporated into the Department of Life Science) of the National Taiwan University and/or Taiwan Museum seems to be lost, it appears necessary to amass new samples, including topotypes, and build up a modern collection of Taiwanese Diplopoda.
The present study is a revision of the millipede genus Chamberlinius, based not only on new material, including some topotypes, but also on a few old types. As a result, all previously described species could be re-assessed, two new species added, and a new synonym established. In addition, the tribe Chamberlinini could be reclassified due to the present revision of the type genus Chamberlinius.
In his Nomenclator, Jeekel (1971, p. 219) corrected Chamberlininae to Chamberliniinae, because family-group names are derived from the stem of the genitive singular of the genus name, which in this case is Chamberlinii, stem Chamberlini-. Similarly, the correct form of Chamberlinini is Chamberliniini. We are keeping Chamberlininae and Chamberlinini because Jeekel's correction was not widely adopted and these names are in prevailing usage (ICZN section 29.5).

Material and methods
New extensive collections of millipedes covering most parts of Taiwan were made between 1986 and 2010, using hand-sorting of the soil and litter. Specimens were preserved in 70% ethanol. External morphology was 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. This material has been shared between the collections of Department of Life Sciences, National Chung-Hsing University (NCHUL), Taiwan; National Museum of Natural Science (NMNS), Taiwan
Type species: Chamberlinius hualienensis Wang, 1956, by original designation. Remarks. Hoffman (1973), when outlining the tribe Chamberlinini, included only two genera therein: Chamberlinius and Riukiupeltis Verhoeff, 1939, with three and one species, respectively. They show very well-developed paraterga resembling those of some Xystodesmidae, both a lobe between ♂ coxae 4 and adenostyles missing, and the gonopods being peculiar in the entire terminal half of the telopodite (distal to the cingulum) representing a single element with several subterminal branches, one of which carries the seminal groove.
Such a description of gonopod structure actually means that, unlike many other tribes of Paradoxosomatidae, which have a long and flagelliform solenomere more or less strongly sheathed/protected by a complex, rather membranous solenophore, the solenomere in the Chamberlinini is a thick, strong and long branch accompanied by a similarly thick, strong and long solenophore branch. The above new diagnosis of Chamberlinius emphasizes its peculiar gonopod conformation. Regrettably, the gonopod conformation of Riukiupeltis still remains to be clarified to properly rediagnose this genus (Jeekel 1968). Jeekel (1968) included the following species in Chamberlinius: C. cristatus (Takakuwa, 1942) from Japan, C. pekuensis (Karsch, 1881) from Japan, Korea, continental China and Taiwan, C. hualienensis (misspelled as hauliensis) Wang, 1956 andC. shengmui Wang, 1957, both latter taxa from Taiwan. Hoffman (1973) later transferred pekuensis into Orthomorphella (see above). As regards C. cristatus, originally described as Orthomorpha cristatus Takakuwa, 1942 from near Tokyo, Japan (Takakuwa 1942), this species badly needs a revision (Jeekel 1968), but the sketch of its gonopod structure alone, showing a flagelliform solenomere and a complex membranous solenophore, prevents its assignment to Chamberlinini altogether. So O. cristatus is here formally ejected from Chamberlinius, but its current generic relationships remain obscure.
As a result, presently only three species can be considered as belonging to this genus (Hoffman 1973, Korsós 2004). Our contribution puts on record two new congeners from Taiwan and establishes the synonymy of one of the older species, altogether providing a revision of Chamberlinius.
Sterna moderately setose in ♂, more sparsely so in ♀; an obvious, short, round ridge supporting spiracles lateral to gonopod aperture (Fig. 8, arrows); each cross-impression with an evident transverse sulcus, but without axial groove. Legs (Fig. 4) without tarsal brushes, long, ca. 1.5 times as long as midbody height in ♂, slightly shorter in ♀. Each ♂ coxa 2 perforated by a vas deferens with a very small pore opening apically.
Distribution: C. hualienensis Wang, 1956 is currently known from Taiwan, as well as on Kyushu Island and on most of the islands of the Ryukyu Archipelago, Japan (Higa and Kishimoto 1986, Yamaguchi et al. 2000, Niijima and Arimura 2002, Nakamura and Korsós 2010. The occurrences of C. hualienensis in Japan are likely to be introduced from Taiwan through human agency. Moreover, it is this species in Japan, but not in Taiwan, that occasionally shows swarming, like repeated massive outbreaks in Okinawa (Higa and Kishimoto 1986) or one at Kagoshima in early winter 2000 which caused serious railway traffic problems (Niijima and Arimura 2002).
Such a distribution pattern vividly reminds of that demonstrated by still another basically Taiwanese paradoxosomatid genus, Aponedyopus Verhoeff, 1939. Of its three currently known species (Chen et al. 2010), only one, the most widespead A. montanus Verhoeff, 1939, has been recorded in the Ryukyus, Japan, but its identity still requires verification (Nakamura and Korsós 2010).
Coloration in alcohol (Fig. 9) light yellow-brown to light brown from head to end of epiproct, as well as from dorsum down to paraterga, sterna and legs; collum and metaterga 2-19 with a slightly infuscate (brown), subtrapeziform band in anterior half (Fig. 11); a lighter or darker anterior part of epiproct; colour pattern similar in both sexes, but ♀ darker; antennae increasingly blackish distally, but tip contrastingly pallid.
Remarks: The sketch of a gonopod of C. shengmui by Wang (1957) showed the solenophore being strongly broadened distad, while the solenomere broadened at midway. Apparently due to these distinctions, this species has since been considered as valid. We have been privileged to examine the specimens of C. shengmui deposited at NMNH (without catalogue number) by Wang, and found the gonopod sketch of the solenomere accompanying the original description (Wang 1957) misleading. Having also reexamined the ♀ holotype and ♂ paratype of C. piceofasciatus, both housed in CAS, and compared them side-by-side with shengmui, we found that these species are identical. Therefore, C. shengmui is a new junior subjective synonym of C. piceofasciatus.

Chamberlinius pessior
Name: To emphasize the lower paraterga. Diagnosis: Closest to C. hualienensis, but differs in being obviously smaller, with paraterga like low ridges (versus higher ridges in C. hualienensis), the pleurosternal carinae are with small caudal teeth on segments 3-10 (versus 3-7(8) in C. hualienensis), legs with tarsal brushes (versus without in C. hualienensis) and the gonopods showing the tip of the solenophore pointed and simple (versus bifid in C. hualienensis).

Key to Chamberlinius species (based mainly on adult males):
1 Dark brown subtrapeziform markings on presulcus halves of metaterga 5-18 divided by a light axial line; solenophore as long as solenomere, parabasal dentiform process of gonopod membranous .

Distribution
The distribution of Chamberlinius species in Taiwan shows that only one species, C. hualienensis, is truly widespread across Taiwan, ranging from lowlands (ca 80 m a.s.l.) to high in the mountains (up to 2,500 m a.s.l.) (Map). Unsurprisingly, it is C. hualienensis that seems to have become introduced to southern Japan. Another species, C. pessior sp. n., has only been encountered at a single midmontane locality at 1,200 m a.s.l., while the remaining two congeners are far more local in distribution, being restricted to high elevations (> 2,000 a.s.l.) both in the northern and central parts of the island. Allopatry or parapatry are prevailing, but syntopic occurrences of two species, one of them usually C. hualienensis, are not too rare, e.g. at Alishan, Hoping etc. To our mind, it seems more logical to surmise that high-montane species in Taiwan tend to be more local in distribution, more inclined to endemism and less liable to introduction than those showing vaster distributions, including lowland habitats. So we are inclined to interpret the presence of the especially common and eurytopic C. hualienensis in Japan as likely introduced from Taiwan.

Reclassification of the Chamberlinini
The subfamily Chamberlininae, proposed by Wang (1957) for Chamberlinius alone, was originally diagnosed as showing a distally "toothed" seminal groove branch (= solenomere). Because this statement was incorrect, Hoffman (1973), who was the first to revise Chamberlinius, downgraded the Chamberlininae to the tribe Chamberlinini and reshaped it as comprising both Chamberlinius and Riukiupeltis. He reformulated the tribe's diagnosis to emphasize that the terminal part of the gonopod lying distally of the postfemoral cingulum represents a single element with several subterminal branches, one of which carries the seminal groove. This diagnosis still holds valid and, based on several recent revisions, including the present one, now allows for another few East to Southeast Asian genera and species to be formally added to the tribe.
Thus, the genus Haplogonosoma Brölemann, 1916, with presumably two valid species (one ranging from central Honshu, Japan to Kurile Island, Russia, the other from Kyushu, Japan to ?Sumatra, Indonesia) shows highly elongate gonopods, in which the slender femorite is supplied with a distinct lateral sulcus. Distally of it, there is a cingulum demarcating a clear-cut geniculation followed by a very long, mesally directed, complex and strongly coiled postfemoral region. The latter is split from its base into a very long, flagelliform, fringed solenomere supported by an undivided, similarly long, slender, membranous lamina (= solenophore). The solenophore carries a pronounced basal outgrowth distodorsally. Golovatch et al. (1995), who revised Haplogonosoma, referred this genus, albeit with hesitations, to the basically Papuan tribe Eustrongylosomatini. At present, however, we believe it is better placed in Chamberlinini.  (Gressitt, 1941) Aponedyopus Verhoeff, 1939, with three species from Taiwan, shows moderately elongate gonopods, in which the femorite is enlarged parabasally, supplied by a distoventral outgrowth, but devoid of a sulcus before a clear-cut cingulum and geniculation. The latter is followed by a long, mesally directed, complex and moderately coiled postfemoral region split from its base into a very long, simple and flagelliform solenomere supported by an only subterminally biramous, similarly long, slender, membranous lamina (= solenophore). The solenophore carries an evident basal outgrowth distoventrally. Chen et al. (2010), in their recent revision of Aponedyopus, have suggested transfer of this genus from the Tonkinosomatini to the Chamberlinini. This transfer is formalized below.
The genus Geniculodesmus Chen, Golovatch & Chang, 2008 has recently been erected (Chen et al. 2008) to accommodate a single species, G. inexpectatus (Attems, 1944), described from Mt Takao, Hachioujishi, Tokyo Prefecture, Japan and also recorded as likely an introduction at Kaohsiung, Taiwan. Its gonopods are elongate, the femorite is slender and devoid of both a distal outgrowth and a sulcus before a clear-cut cingulum and geniculation. The latter is followed by a long, mesally directed, rather simple and moderately coiled postfemoral region split from its base into a long, simple and flagelliform solenomere supported by an only subterminally bifid, similarly long, slender, membranous lamina (= solenophore). The solenophore carries a basal, caudally directed outgrowth.
As regards Riukiupeltis Verhoeff, 1939, its scope and diagnosis remain rather obscure. The type-species R. yamashinai Verhoeff, 1939, from the Ryukyus, Japan has been described as showing the gonofemorite long and slender, devoid of both a sulcus and a distal outgrowth before a clear-cut cingulum and geniculation. The postfemoral part is very long, slender, probably directed mesad, moderately coiled, devoid of any basal outgrowth, gradually attenuating towards a pointed tip. A solenomere is only depicted as likely broken off, branching off from the distal half of the solenophore (Verhoeff 1939).
Such a conformation seemed so improbable that Jeekel (1968) doubted it altogether, especially as regards the extent and position of the solenomere. The more so as in another species, R. falcatus (Attems, 1953), from Laos and Vietnam, he also formally placed in Riukiupeltis the gonopods show a condition readily reminding of that observed in Chamberlinius: the solenomere is not flagelliform, but thick and long, gradually attenuating towards a pointed tip (Figs 58, 59). Yet, pending a revision of the type species R. yamashinai, it appears impossible to unequivocally state that Riukiupeltis is indeed characterized by the gonopod postfemoral portion being normal, non-split, devoid of any additional solenophore, being represented instead solely by a thick, long and helicoidally twisted solenomere.
To accommodate the above five genera into the tribe Chamberlinini, and to properly rediagnose it, only a few slight amendments to Hoffman's (1973) definition are necessary, as follows.
Tribe Chamberlinini Wang, 1957 Diagnosis: Gonopod femorite usually long and slender, only rarely somewhat expanded parabasally and supplied with a distal outgrowth (Aponedyopus) or carrying a lateral sulcus (Haplogonosoma) before a clear-cut cingulum and geniculation. Postfemoral portion directed mesad, long, slender, more or less coiled, either entirely a thick solenomere devoid of any outgrowths (Riukiupeltis) or split at base into a thick (Chamberlinius) or thin, long, truly flagelliform solenomere and a membranous, more or less complex solenophore (remaining three genera).