New cave-dwelling species of Tomoceridae from China, with a study on the pattern of mesothoracic bothriotricha in Tomocerinae (Collembola, Entomobryomorpha)

Abstract Two new troglobitic species of Tomoceridae are described from Guizhou and Guangxi provinces, China. Tomocerus tiani sp. n. resembles Tomocerus kinoshitai Yosii, 1954, Tomocerus caecus Yu & Deharveng, 2015 and Tomocerus similis Chen & Ma, 1997 but differs from them mainly in the body colour, the cephalic chaetotaxy and the number of manubrial pseudopores. Monodontocerus cinereus sp. n. is similar to Monodontocerus mulunensis Yu, Deharveng & Zhang, 2014 but is different from the latter in the body colour, the length of antennae, the number of ungual teeth and the chaetotaxy on Abd. III and Abd. IV. Special remarks are made on the mesothoracic bothriotricha in Tomocerinae.


Introduction
Since the discovery of Tritomurus scutellatus Frauenfeld, 1854 in Slovenian caves, troglobitic Tomoceridae have frequently been reported in Asia, Europe and North America. Besides the troglobitic genera Tritomurus Frauenfeld, 1854 andLethemurus Yosii, 1970, some other main groups of Tomoceridae have also been found with cave dwellers, and several genera, i.e. Monodontocerus Yosii, 1955, Plutomurus Yosii, 1956 and Aphaenomurus Yosii, 1956 have mainly or usually been found in caves. However, despite of several highly troglomorphic species, for example Tritomurus falcifer Cassagnau, 1958 and Tritomurus veles Lukić, Houssin and Deharveng, 2010 which are both eyeless and have very elongated antennae and claws, other cave tomocerids do not exhibit distinct troglomorphic characters: most of them have short to moderate antennae, normal or slightly elongated claws and a full set of 6+6 eyes for Tomocerinae, only the sizes of eyes are usually smaller than those of the edaphic species, and the tenent hair is often pointed as in other cave Collembola.
Four cave tomocerids have previously been reported in China, including Tomocerus caecus Yu & Deharveng, 2015, Monodontocerus absens Yu, Deharveng & Zhang, 2014 and Monodontocerus mulunensis Yu, Deharveng & Zhang, 2014 from Guangxi Province, and Monodontocerus trigrandis Yu, Deharveng & Zhang, 2014 from Hunan Province. The present paper reports two new species of Tomoceridae discovered in caves in the south-west karst regions of China. Both of the new species have small eyes and pointed tenent hair, but neither of them is highly troglomorphic. Tomocerus tiani sp. n. bears some unusual characters, including the single mesothoracic bothriotrichum, which lead to a comprehensive examination of the different genera of Tomocerinae.

Materials and methods
Specimens were collected with aspirators and preserved in 99% ethanol. For detailed morphological study, specimens were cleared in Nesbitt's fluid and mounted in Marc André II solution. For some specimens the furca, the ventral tube and the legs were cut off from the trunk and mounted separately for detailed observation. The slide-mounted specimens were studied using a Nikon Ni microscope. Photos were taken using Nikon DS-Fi1 cameras mounted respectively on Nikon SMZ1000 stereomicroscope and Nikon Ni microscope. Fjellberg (2007) is followed for maxillary lamellae numbering, Yu et al. (2014) for the pattern of cephalic dorsal chaetotaxy and Christiansen (1964) for body macrochaetotaxy. The description of the body chaetotaxy refers to one side only since in most case it is symmetric. The exact morphology of each chaeta was unclear due to shedding. The dental spines formula follows that of Folsom (1913), in which the dental spines are arranged from basal to distal, with a slash indicating the separation between basal and medial subsegments and the Roman numerals referring to spines that are noticeably larger. If not mentioned specially, all descriptions are based on fully developed individuals. Type-specimens. Holotype male (labelled 15cave15-1) and paratype juvenile (labelled 15cave15-2) on slide. Deposited in NJAU.

Abbreviations
Diagnosis. Species similar to Tomocerus kinoshitai Yosii, 1954, with short antennae, multi-furcated dental spines and apically curved mucro. Body length approximately 3.0 mm, with purplish grey pigment all over; antennae approximately half as long as body; eyes small; terminal hair of maxillary outer lobe with a small basal denticle; Th. II with only one bothriotrichum; tenent hairs pointed; unguis with two teeth, baso-internal ridges at approximately 1/2 distance from base; manubrium with 12-17 pseudopores on each side; dental spines formula as 4/2, II; dens dorsally with only a few feather-like chaetae; mucro without intermediate tooth. Cave-dwelling species.
Description. Body length 2.9 mm. Body with uniform purplish grey pigment and unpigmented patches, appendages paler. Eye patches black (Fig. 1A). Types of scales and chaetae typical for Tomocerinae.
Trochantero-femoral organ with 1, 1 small slender chaetae (Fig. 2E). Front, middle and hind tibiotarsus ventrally with 0, 0, 2 pointed spine-like chaetae (Fig. 2F). Each tibiotarsus with a distal whorl of 11 chaetae, ventral six as ordinary chaetae, dorsal five modified: tenent hair thin and pointed, approximately 0.33 times as long as inner edge of unguis; two accessory chaetae small, longer than pretarsal chaetae; two guard chaetae of same morphology and size as tenent hair. Unguis slender, with baso-internal ridges at approximately 1/2 distance from base; lateral teeth pointed, of moderate size. Inner edge of unguis with one basal and one central minute teeth. Unguiculus rather slender, approximately 0.5-0.72 times as long as unguis, its inner edge with one corner tooth. Pretarsus chaetae 1+1 (Fig. 2G). Anterior face of ventral tube with scales, posterior face and lateral flaps unscaled, anterior face with ca. 25 chaetae on each side, posterior face with ca. 30 chaetae, each lateral flap with ca. 15 chaetae. Rami of tenaculum with 4+4 teeth, anterior face with one chaeta and without scale (Fig. 3A). Length ratio of furca segments as manubrium : dens : mucro = 2.5 : 3.6-3.7 : 1.0. Manubrium ventrally scaled, without chaetae, laterally with large round scales and 7-9 strong chaetae; dorsal chaetal stripe with ca. 200 chaetae of different sizes, including 2+2 pointed prominent chaetae; inner side of chaetal stripe with several scales; pseudopores 12-17 on each side (Fig. 3B); external corner chaeta as a microchaeta (Fig. 3C). Dens basally with a pointed prominent dorsal chaeta, without large modified inner scale or strong outer chaetae. Dental spines formula as 4/2, II; spines with moderate to large sized denticles around basal half (Fig.  3D). Dens dorsally with ordinary chaetae, swollen spine-like ciliated chaetae and a few feather-like chaetae (Figs 1B, 3E), ventrally with scales and several apical chaetae. Mucro elongated, distally curved and multi-setaceous; both basal teeth with proximal lamellae, outer tooth with a toothlet; apical and subapical tooth subequal; two dorsal lamellae running from subapical tooth, outer lamella ending in inner basal tooth, inner lamella ending at base of mucro; without intermediate teeth (Fig. 3F) Chen & Ma, 1997 (type materials) in the length of antennae, the general pattern of chaetotaxy on Th. II, the number and position of spine-like tibiotarsal inner chaetae, the size of external corner chaeta on the manubrium, the type and general arrangement of dental spines and the shape of mucro, but it is clearly different from the first species in having eyes and pigment, and is different from the other two species mainly in the body colour, the cephalic chaetotaxy, the sharply pointed tibiotarsal strong inner chaetae and tenent hair, and the more slender unguiculus; besides, with similar body size, Tomocerus tiani sp. n. has more manubrial pseudopores than the three known species. The small basal denticle of the terminal hair of maxillary outer lobe is so far unique to Tomocerus tiani sp. n. and is useful for diagnosis if dissected and exposed carefully. The baso-internal ridges of unguis are located at approximately 1/2 distance from the base in Tomocerus tiani sp. n., whereas in most other species the distance between the ridges and the base is only 1/3 or less. The dorsal dental chaetae in Tomocerus tiani sp. n. is also characteristic, that the dense stripes of feather-like chaetae in most other tomocerids are almost replaced by ordinary chaetae and swollen serrated chaetae, leaving only a few feather-like ones. Similar condition was also reported in Tomocerus kinoshitai and Tomocerus caecus that some spine-like chaetae are present on dens (Yosii 1967, Yu andDeharveng 2015).
The juvenile specimen is almost identical to the adult in most characters, including the macrochaetotaxy, the number of teeth on claws and the dental spines formula. However, some characters on manubrium are distinctly different between juvenile and adult. In the juvenile specimen, there are ca. 80 dorsal chaetae on each side of manubrium, the number of pseudopores is only 4-5 on each side, and the external corner chaeta is as large as a mesochaeta in the dorsal chaetal stripe. These differences provide interesting information in the postembryonic development of manubrium in Tomocerinae, and indicate these characters are not suitable for the identification of immature specimens at different instars.  Type-specimens. Holotype male (labelled 15cave6-1) and two paratypes female (labelled 15cave6-2 and -3) on slides, one paratype (labelled 15cave6) in alcohol. Deposited in NJAU.
Etymology. Named for its light grey body colour, from the Latin cinereus, meaning ash-coloured.
Remarks. Within the genus, Monodontocerus cinereus sp. n. is more similar to Monodontocerus mulunensis Yu, Deharveng & Zhang, 2014 in the cephalic chaetotaxy, the pointed tenent hair and the number of mucronal intermediate teeth, but can be distinguished from the latter by having longer antennae, fewer teeth on unguis and more macrochaetae on Abd. III and Abd. IV. In alcohol the new species can be identified from other known species of Monodontocerus by the grey body colour.

Discussion
Tomocerus tiani sp. n. has only one bothriotrichum on Th. II, while there are two in most other Tomocerus species observed previously. The pattern of bothriotricha is significant for the taxonomy and phylogeny of Collembola (Szeptycki 1979, Soto-Adames et al. 2008. However, the exact pattern of mesothoracic bothriotricha in Tomocerinae has not been resolved though the number of them was commonly reported as either two (Yosii 1956, 1967, Fjellberg 2007 or one (Christiansen 1964, Chen andMa 1997).
There are two main obstacles to determining the pattern of mesothoracic bothriotricha in Tomocerinae. Firstly, these long, thin and ciliated chaetae, as well as the macrochaetae, are easily lost during specimen collection and slide preparation, leaving only the sockets. The sockets of bothriotricha are usually characteristic for their round shape and small size, but sometimes can still be confused with the sockets of macro-or mesochaetae. Secondly, in case of two bothriotricha present they are arranged transversely, and the outer one is usually near the lateral margin of the tergum, thus could possibly be omitted when the margin is wrinkled or damaged. To avoid those disadvantages, we examined the specimens with almost intact coating of chaetae, most of which are pre-molting specimens with new chaetae under the old cuticle.
In the observed species, there are three main patterns of mesothoracic bothriotricha, which appear to be relevant to the generic division except for several species of Tomocerus and Tomocerina. Pattern A: all species of Tomocerus ocreatus complex , Tomocerus nigrus Sun, Liang & Huang, 2006 and Tomocerus jilinensis Ma, 2011 have two mesothoracic bothriotricha, and the socket of the inner one is usually larger, thus is more similar to that of a macrochaeta (Fig. 6A) Ma, Chen & Christiansen, 2003 have only one bothriotrichum at approximately the place of the inner bothriotrichum in other Tomocerus, and a macrochaeta is present instead of the outer bothriotrichum (Fig. 6B, C, D). In the first two species this macrochaeta is longer and apically more tapered than the adjacent macrochaetae (Fig. 6B); in Tomocerina annamitica this macrochaeta is slender ( Fig. 6C), and its socket is very similar to that of a bothriotrichum; in the last three species the macrochaeta is large, rather elongated, pointed, and is more ciliated than other ordinary chaetae, forming a bothriotricha-like macrochaeta (Fig. 6D). Pattern C: in Pogonognathellus heterochros Wang, Yu & Zhang, 2013, Pogonognathellus mai Wang, Yu & Zhang, 2013 and Pogonognathellus sp. from France, there is only one bothriotrichum, and no distinctly special macro-or mesochaetae is present nearby (Fig. 6E).
This study has revealed several distinct patterns of mesothoracic bothriotricha in Tomocerinae, and has proved their taxonomic importance. However, since our study covered mostly Asian edaphic species, we have probably not exhausted the variability of this character among Tomocerinae. The exact pattern in Monodontocerus has not been successfully determined because of the lack of specimen with intact bothriotricha, though the sockets indicate pattern A more possible. For Pogonognathellus more examination is required since several species were described having two mesothoracic bothriotricha (Fjellberg 2007, Park et al. 2011). On the other hand, given the thoracic bothriotricha do not exist in the primary chaetotaxy (Szeptycki 1972, they could either be secondary elements or be transformed from certain primary chaetae. Tracing the postembryonic development of chaetotaxy will help to evaluate the true significance of thoracic bothriotricha for taxonomy, while study on the homology of chaetae will be a next step to review the current taxonomic system of Tomocerinae.