Occurrence of the millipede genus Tonkinosoma Jeekel, 1953 in China, with the description of the first presumed troglobitic species of this genus (Diplopoda, Polydesmida, Paradoxosomatidae)

Abstract The genus Tonkinosoma Jeekel, 1953 has hitherto been known to contain only two species, both from northern Vietnam. T. flexipes Jeekel, 1953, the type species of the genus, is recorded from Guangxi, southern China, for the first time. T. tiani sp. n., a presumed troglobite, is described from caves in Guizhou, southwestern China. A key is presented to all three species of the genus.


Introduction
The millipede genus Tonkinosoma Jeekel, 1953 was originally proposed to encompass only a single species, T. flexipes Jeekel, 1953, from northern Vietnam (Jeekel 1953). The tribe Tonkinosomatini Jeekel, 1968 was erected to comprise not only Tonkinosoma, but also a few other genera (Jeekel 1968). However, this tribe has since been merged first with Sulciferini Attems, 1898 (cf. Golovatch 2014) and, later, with Chamberliniini Wang, 1956(cf. Golovatch 2015a. Nguyen (2011), when reviewing Tonkinosomatini in the scope of the fauna of Vietnam, described a second species from northern Vietnam: T. jeekeli Nguyen, 2011. At present, Tonkinosoma can be diagnosed as a genus of the Himalayan and southeast Asian tribe Chamberliniini characterized by the gonopods which show a postfemoral part demarcated basally by an indistinct (T. flexipes) or distinct (T. jeekeli) geniculation cingulum and distally at least by a lateral sulcus. Unlike the other contribal genera, the femorite is long and slender and, like the postfemoral part, it is devoid of outgrowths, the seminal groove runs all along the mesal face of the femorite, the solenomere is long and flagelliform, and the solenophore is a long, hyaline, folded lobe that shows a lamina medialis and a lamina lateralis, both sheathing the solenomere. As in most Chamberliniini, both solenomere and solenophore are usually subcircular (Chen et al. 2010;Nguyen and Korsós 2011;Golovatch 2014). Golovatch (2014) questioned the attribution of T. jeekeli to Tonkinosoma, but now we rather believe that he somewhat misinterpreted the gonopodal structure of T. flexipes as described and illustrated by Jeekel (1953). Instead we follow Nguyen (2011) and consider both T. flexipes and T. jeekeli to represent congeners.
Prompted by the discovery of both T. flexipes and a third, new species of Tonkinosoma in southern China, the latter species the first to be found in caves, their descriptions are provided below. We also provide a key to all three presently known species of this genus.

Material and methods
All specimens used in this study were collected in southern China and preserved in 95% ethanol. Most of the type and non-type material is deposited in the Zoological Collection of the South China Agricultural University, Guangzhou, China (SCAU), with several samples also donated to the Zoological Museum, Moscow State University, Moscow, Russia (ZMUM), as indicated below.
Observation and dissections were performed using a Leica S8 APO stereo microscope. The line drawings were prepared with a Leica MZ125 microscope and a camera lucida attached to the microscope.
Photographs were taken with a Keyence VHX-5000 digital microscope, and further edited using Adobe Photoshop CS5. A large, median, subquadrate process between ♂ coxae 4 and two small, independent tuberculations between ♂ coxae 5 ( Remarks. This is the type species of Tonkinosoma hitherto known only from a highly detailed original description, based on the male holotype and two paratypes, one male and one female, all from Mt Manson, Langson Province, northern Vietnam (Jeekel 1953). Above is only the second record of T. flexipes, a species new to the fauna of China, but this is hardly too surprising because it comes from a place quite close to the border with northern Vietnam. The new samples almost fully agree with the original description (Jeekel 1953), but our material is remarkably smaller (19-28 mm vs.  37-47 mm). The habitus and gonopod structure (Figs 1-3) are illustrated to document the species' identity. Among the main diagnostic characters of T. flexipes, the following seem to be especially noteworthy to complement the only available description: integument strongly shining; metazonae with several longitudinal striae above paraterga (Fig. 2C); pleurosternal carinae present on segments 2-4 in both sexes (Fig. 2C); a large, median, subquadrate process between ♂ coxae 4 and two small, independent tuberculations between ♂ coxae 5 (Fig. 2F); legs ca 1.6 (♂) or 1.2 (♀) times as long as midbody height, tarsal brushes present on all ♂ legs (Fig. 2I); gonopodal postfemoral part only indistinctly demarcated, lamina lateralis well-developed only in the proximal part of the solenophore (Fig. 3). Diagnosis. This new species differs from its congeners in showing a largely unpigmented body. It seems to be especially similar to T. jeekeli on account of the particularly elongate and subcircular solenophore and solenomere, but differs by the strongly developed pleurosternal carinae present until segment 17 in both sexes, by an evident, subtrapeziform process between ♂ coxae 4, and the gonopod with a small and sharp tooth near the base of the solenomere.
Gonopods (Figs 6D,7,8) simple. Coxite relatively short, about half as long as telopodite, poorly setose both distodorsally and distoventrally. Prefemoral portion densely setose as usual, about as long as coxite. Femorite (fe) long and slender, slightly curved mesally and faintly enlarged distally. An obvious demarcation sulcus (s) both laterally and dorsally between fe and a postfemoral portion (pf). Solenophore (sph) clearly coiled, circular, both lamina medialis (lm) and lamina lateralis (ll) well-developed and nearly entirely sheathing a similarly long and free solenomere (sl). Seminal groove (sg) running entirely on mesal side of femorite before moving onto sl, with a very small, sharp, mesal tooth (t) on pf near sl base.
Remarks. The karstic Ganhan Dong cave where the holotype was taken is about 300 m long. All material was collected in areas of complete darkness.
Based on the largely unpigmented integument, the long legs (2.5 (♂) or 2.0 (♀) vs. 1.6 (♂) or 1.2 (♀) times as long as midbody height in T. flexipes) and the cave habitat, this species seems to be a troglobite.

Discussion
The above record of T. tiani sp. n. in caves in southern China is remarkable at least in two ways. Firstly, the huge family Paradoxosomatidae only rarely occurs in caves, with only few presumably troglobitic species. The only exceptions are in the large genus Desmoxytes Chamberlin, 1923, which is very common both in epigean and subterranean environments across southeast Asia and China, and in the small genus Piccola Attems, 1953, with a few epigean species in Vietnam and Laos, and a single troglo-bitic one from Guangxi, China (Liu and Tian 2013;Golovatch 2015b). Secondly, biogeographically the situation concerning the distribution pattern of Tonkinosoma strongly resembles that not only of Piccola, but of still another millipede genus, i.e., Pacidesmus Golovatch, 1991 (Polydesmida, Polydesmidae). The latter genus has one high-mountain species in northern Thailand and a further eight, all presumed troglobites, in southern China (Golovatch and Geoffroy 2014).
Further research on cave millipedes of China will definitely reveal not only new interesting taxa, but more cases of remarkable distribution patterns.