A new and cryptic species of Lissodesmus Chamberlin, 1920 (Diplopoda, Polydesmida, Dalodesmidae) from Tasmania, Australia

Abstract Lissodesmuspiscatorsp. nov. differs from the 30 previously described Lissodesmus species in the form of the femoral process of the gonopod telopodite, which is tripartite with an erect distal branch and two posteromedially curving basal branches. Despite careful searching, the new species has only been collected by pitfall trapping and may have a very small range in the northwest corner of the Central Plateau in Tasmania, Australia.


Introduction
The genus Lissodesmus Chamberlin, 1920 currently includes 19 species in Tasmania and 11 species in Victoria; see Mesibov (2018) for a genus synonymy and list of species. Although the Tasmanian Lissodesmus fauna has been well sampled over many years (Fig. 1A), a narrow-range alpine species was only first collected in 2017 (Mesibov 2018) and additional narrow-range species probably remain to be discovered.
The new Lissodesmus species described in this paper might represent another category of undescribed Tasmanian millipedes: possibly more widely distributed, but unusually cryptic. It was pitfall-trapped at a single location in the summers of 2017, 2018 and 2019, but has not yet been collected by hand in the pitfall area, despite careful searching.

Pitfall trapping
Pitfall traps were set and emptied by Michael Driessen of the Department of Primary Industries, Parks, Water and Environment (DPIPWE), Tasmania. The trapping was part of a wildfire recovery study following a fire that burned from January to March 2016 and covered ca. 26,000 ha south and west of Lake Mackenzie, a hydroelectric impoundment at the northwest corner of Tasmania's Central Plateau (Natural Values Conservation Branch 2017). Post-fire monitoring studies of flora and fauna were initiated shortly after the fire had been brought under control.
The pitfall traps were set on Ritters Plain (Fig. 2), ca. 2.5 km west of Lake Mackenzie and ca. 1 km south of the Fisher River gorge. Each trap was a 225 ml plastic cup placed with its rim flush with the ground surface in a short section of 75 mm PVC pipe fitted snugly in a hole drilled with an auger. The traps were placed in each of three burn categories: unburned vegetation and unburned peat, burned vegetation and unburned peat (lightly burned) and burned vegetation and burned peat (deeply burned). Ten traps were irregularly placed a minimum of 5 m apart in each category, reflecting the patchy nature of the burn. Each trap cup was filled with 100 ml of 70% ethanol over a few ml of glycerol, covered with a rain shield and left for 14 days in the late austral summer: 20 February -6 March 2017, 22 February -8 March 2018 and 20 February -6 March 2019.  Tasmania  Millipedes and other invertebrates in the traps were sorted for DPIPWE by Kevin Bonham, who sent specimens of an unfamiliar Lissodesmus to the author for identification in May 2018.

Millipede searches
I searched the Ritters Plain pitfall area and the nearby grassy sedgeland and subalpine forest unsuccessfully for fresh material of the new Lissodesmus on 8 December 2018 and on 20 February, 11 March and 28 March 2019, for a total of ca. 8.5 hours over ca. 20 ha. I collected representative specimens of other millipede species on each visit. Microhabitats examined included woody litter, bark litter, leaf litter, grass and sedge turf, and the spaces under stones and under prostrate (rock-hugging) shrubs. I also excavated small deposits of peaty soil underlying elevated Sphagnum moss mounds on the south side of the Plain.

Specimen preparation
All specimens of the new Lissodesmus species are stored in 80% ethanol in the Queen Victoria Museum and Art Gallery. The paratype male was briefly treated with vinegar to reduce its stiffness. Specimens were examined and measured using a Nikon SMZ800 binocular dissecting microscope. Focus-stacks of colour images were manually generated using a Canon EOS 1000D digital SLR camera mounted on the Nikon SMZ800 fitted with a beam splitter, then processed with Zerene Stacker 1.04 software. The gonopods of a male from 2018 pitfall trapping were cleared in 80% lactic acid, temporarily mounted in a 1:1 glycerol:water mixture and imaged using an eyepiece video camera mounted on an Amscope binocular microscope. Preliminary drawings were traced from printed copies of the images, then corrected by reference to the actual gonopods. Figures were composed using GIMP 2.8 and the map in Fig. 1A with QGIS 2.14.
Specimen locality data are provided in Supplement material 1 in Darwin Core format. Pitfall trap locations were provided by Michael Driessen as a site map based on a georegistered aerial photograph (see Fig. 2A). Locations of the 10 traps yielding the new Lissodesmus species have been spatially summarised as -41.6824 146.3524 ±75 m (WGS84 datum).

Millipede searches
As mentioned in the Materials and methods section, I found no Lissodesmus specimens in or near the pitfall area. However, I had little difficulty finding the four other trapped species, as well as four more: Gasterogramma psi Jeekel, 1982 and "M20" (both Polydesmida: Dalodesmidae, the latter undescribed but recorded elsewhere in northwest Tasmania), an unidentified Procyliosoma sp. (Sphaerotheriida, Procyliosomatidae) and the undescribed but well-recorded siphonotid "AcuMes" (Polyzoniida: Siphonotidae). Paratype. AUSTRALIA • male; same data as for holotype; dissected; QVMAG: QVM:2019:23:0016.

Order Polydesmida Pocock
Other material. 5 males, 4 females and 1 stadium 7 female, same locality as holotype; see Supplement material 1 for details.
Diagnosis. Distinguished from all other known Lissodesmus species by the form of the femoral process on the gonopod telopodite: the process has an erect, flattened, bluntly toothed distal branch and two large basal branches curving posteromedially across the posterior face of the telopodite.
Female closely resembling male but stouter. Genital aperture with posterior margin gently convex medially; cyphopods not examined.
Name. Latin piscator, fisherman, noun in apposition, for the type locality in the Fisher River catchment.
Distribution. So far known only from Ritters Plain near Lake Mackenzie in northwest Tasmania (Figs 1B, 2). The Plain has a habitat area of ca. 100 ha and a known occupied area of less than 1 ha.

Discussion
In what microhabitats are L. piscator sp. nov. living?
Ritters Plain is almost entirely treeless (Fig. 2). Most of the Plain is covered by a tightly intergrown turf of grass and sedge species overlying a dense, fibrous, anaerobic peat, mixed with patches of slightly elevated Sphagnum bog (Department of Primary Industries, Parks, Water & Environment 2016, Natural Values Conservation Branch 2017. Annual rainfall at nearby Lake Mackenzie is ca. 2100 mm and snow lies on the Plain during the winter months. For much of the year the soil is saturated with water. The pitfall area was probably even wetter before a canal ( Fig. 2A) was constructed in the 1960s on the uphill side of the Plain to capture additional water for the Fisher River hydroelectric project.
There are no above-ground shelters near the pitfall traps in which L. piscator sp. nov. was found, so the population is probably living underground, either in crevices in the root-filled turf or in cavities among the periglacially shattered rock fragments that cover much of this portion of the Central Plateau and underlie the peaty soil. Adults might be expected to come to the surface to mate and disperse during late summer -when the pitfall trapping was carried out -and 11 of the 12 specimens trapped were adults.
The densest populations of millipedes I found by searching on Ritters Plain were in surface peat associated with elevated Sphagnum moss mounds, such as the ones surrounding the lower parts of Richea scoparia stems just to the southwest of the pitfall area (s in Fig. 2A). The foliage and fine branches of the Richea were burned in the 2016 fire, but many of the lower stems buried in moss were still alive in 2019. The moss mounds themselves, with and without Richea, were largely unburned. In the peat close to the moss I frequently found G. psi and Procyliosoma sp., and occasionally small groups of P. monticolus and A. fossuliger. I was surprised not to find L. piscator sp. nov. in the peat as well, since Lissodesmus spp. cohabit with Gasterogramma spp. elsewhere in Tasmania.
portion of Ritters Plain that coincidentally was sampled with pitfalls in the post-fire recovery study. I suspect that it also occurs nearby in the voids in the periglacial scree deposits that cover slopes with boulder-sized rocks in the northwest corner of the Central Plateau. Sampling in the screes would be even more difficult than setting pitfall traps over a wide area, and the true distribution and conservation status of L. piscator sp. nov. are likely to remain indeterminate.