A new millipede genus and a new species of Asphalidesmus Silvestri, 1910 from southern Tasmania 55 A new millipede genus and a new species of Asphalidesmus Silvestri, 1910 (Diplopoda, Polydesmida, Dalodesmidea) from southern Tasmania, Australia

Noteremus summus gen. n., sp. n. occurs at 1100-1300 m on the summit of Mt Weld, southern Tasmania, while its congener N. infi mus sp. n. is troglobitic in caves in the Junee-Florentine karst, 30-40 km to the northwest. Like species of Paredrodesmus Mesibov, 2003 and Procophorella Mesibov, 2003, Noteremus spp. have a head + 19 rings, no sphaerotrichomes and pore formula 5, 7-18, and are not assigned to family within the suborder Dalodesmidea. Asphalidesmus golovatchi sp. n. occurs in caves and in forest litter in far southern Tasmania, and the adults have paramedian and median tergal projections. Asphalidesmus Silvestri, 1910 is removed from Polydesmidea, Haplodesmidae and placed in Dalodesmidea without family assignment.


Introduction
Tasmania has a rich and entirely endemic fauna of dalodesmidean millipedes, i.e. members of order Polydesmida, suborder Dalodesmidea.Sixty species have so far been named (Mesibov 2009) and at least 50 more are in collections awaiting description.In this paper I describe three new species from Tasmania's south, a mountainous district which has not yet been carefully surveyed for rare and geographically restricted invertebrates.Two of the species, a troglomorphic cave-dweller and a non-troglomorphic surface-dweller, are placed in a new genus.Th e new cave-dweller is the third troglobitic millipede known from Tasmania.Th e other known troglobites, Atalopharetra clarkei Mesibov, 2005 andA. eberhardi Mesibov, 2005, are likewise in a southern Tasmanian genus with surface-dwelling congeners (Mesibov 2005).Th e third species described here occurs both in caves and in forests and is placed in Asphalidesmus Silvestri, 1910.

Methods
"Male" and "female" in the text refer to stadium VII individuals unless otherwise indicated.All specimens are stored in 80% ethanol in the Queen Victoria Museum and Art Gallery, Launceston, Tasmania, Australia (QVM).Gonopods were cleared and temporarily mounted in 60% lactic acid for optical microscopy; other body parts were temporarily mounted in a 1:1 glycerine-water mixture.Preliminary drawings on graph paper were made using an eyepiece grid at 64X or 160X.SEM images were acquired digitally using an FEI Quanta 600 operated in high-vacuum mode; alcohol-preserved body parts were air-dried before sputter-coating with gold.
Where given, geographic coordinates (latitude, longitude) are based on the WGS84 datum.Cave locations are treated specially.Of the ca 4000 known caves in Tasmania, fewer than 200 are listed in the offi cial Tasmanian Government gazetteer of named places.Th e locations of other caves are kept secret by caving enthusiasts.Cave locations are only revealed to members of recreational caving clubs and to a small number of Tasmanian Government employees.Both the clubs and the Government actively discourage publication of exact cave locations.For this reason I provide here only approximate geographic coordinates for cave locations, with an uncertainty of ±1 km.
Readily distinguished from Paredrodesmus by the presence of paranota as large lateral swellings on posterior segments, and from Procophorella by the absence of a narrow, well-defi ned, upwardly concave groove on the paranotal margin.
Remarks.Noteremus species are dalodesmidean in that the small, weakly joined gonocoxae are completely withdrawn into the aperture.However, the absence of sphaerotrichomes sets the genus apart from Dalodesmidae s. str.I therefore place Noteremus in the suborder Dalodesmidea without assigning it to a family.I did the same (Mesibov 2003) with the Tasmanian H+19 genera Paredrodesmus Mesibov, 2003 andProcophorella Mesibov, 2003, which like Noteremus have the unusual pore formula 5, 7-18.Th ere are other similarities.In all species of all three genera, the ventral pair of spinneret setae is further apart than the dorsal pair (Figs 4A-4D; type species of Paredrodesmus and Procophorella illustrated as examples).In Noteremus, Procophorella and four of the six described Paredrodesmus species the solenomere is a short, acuminate process near the telopodite apex, and in Noteremus and fi ve Paredrodesmus species the telopodite bears clusters of stout, pointed, rod-like structures.It thus seems likely that Noteremus, Paredrodesmus and Procophorella constitute a natural group, but I am reluctant to erect for them a new family or subfamily at this time (see discussion below on Asphalidesmus).
Clusters of rod-like structures are present on gonopods in the dalodesmid genus Icosidesmus Humbert & de Saussure, 1869.Attems (1940) referred to the structures as "starker Stifte", which I translate as stout pegs or pins, while Johns (1964) called them "long, stout setae".Jeekel (2006) examined two New Zealand Icosidesmus species and named the structures "bacilli" (rods) and the clusters "bacillaries".A similar cluster on the gonopod of Tasmaniosoma armatum Verhoeff , 1936 was called "eine starke Borstengruppe" by Verhoeff (1936), i.e. a stout group of bristles.If these structures are homologous, it is interesting that they can appear in diff erent places on the telopodite, i.e. at mid-height in parallel clusters in Icosidesmus and Tasmaniosoma, in upright parallel clusters in Noteremus, and in sub-apical, sometimes fan-like clusters in several Paredrodesmus species.

A
A A swellings at ring mid-height, tapering towards posterior corner.Ozopore small and round, opening dorsolaterally at about half ring height, close to posterior paranotal corner; pore formula 5, 7-18.Spiracle small, round, recessed, opening on short, widerimmed elevation; anterior spiracle on diplosegments opening just above and anterior to anterior leg base, posterior spiracle above and about midway between leg bases.Sternites longer than wide, very sparsely setose, longitudinal and transverse impressions well-defi ned.Pre-anal ring with a few setae; epiproct broadly rounded, only slightly extending past anal valves; hypoproct trapezoidal.Ventral spinnerets ca 2.5X further apart than dorsal spinnerets (Fig. 4A); spinneret seta set in thin, closely fi tting sheath with irregular margin.Anterior legs (Fig. 2A) a little swollen with prefemur and femur somewhat expanded dorsally, tarsus straight, claw small; relative podomere lengths tarsus>femur>prefemur>(postfemur, tibia).No sphaerotrichomes; brush setae on prefemur, femur, postfemur, tibia; brush setae unbranched with tapered tips.Gonopore small, round, on distomedial projection of leg 2 coxa.Bases of legpairs 5, 6, 7 separated to accommodate retracted gonopods, legpair 5 bases less so; small, paired, conical projections on sternal portion of legpair 4 bases, each projection tipped with small brush of setae (Fig. 4F).Gonopod aperture about one-third prozonite width, rhomboid with long anterior edge close to anterior prozonite margin; posterior rim of aperture slightly raised.Gonocoxae small, short, tapering distally, concave mediobasally, weakly joined mediodistally.Cannula prominent, inserting in shallow depression on basal surface.Telopodites separate, reaching to legpair 4 bases when retracted.Telopodite (Figs 5A, 5B, 6A) thin, straight, unbranched, with fl ared base and equally expanded apex.Solenomere a short, acuminate process on medial side of telopodite apex.Prostatic groove bending anteriorly in fl ared telopodite base, then posterodistally and running more or less directly to solenomere on medial side of telopodite.Telopodite with numerous short, very fi ne setae at base near cannula insertion; a few longer setae on posterolateral surface close to base; and two closely packed groups of pointed, rod-like structures at or near the apex: a distal row extending distolaterally to a small cluster, and two separate posteromedial clusters at either end of a short, low posteriormedial ridge near the telopodite apex.Female (Fig. 1A) larger than male, anterior legs not swollen; posterior margin of epigynum produced medially as narrow, round-tipped projection, just reaching beyond leg 2 coxae ventrally; cyphopods not examined.

A C B D
Distribution and habitat.So far known only from grassland, scrub and subalpine woodland on the summit of Mt Weld, southeast Tasmania, from 1100 to 1300 m elevation (Fig. 9).
Remarks.N. summus is the largest H+19 polydesmidan in Tasmania.Th e next largest species, an undescribed species of Tasmaniosoma Verhoeff , 1936, has a maximum diameter in females of less than 1.5 mm, compared to 2.2 mm in N. summus.All specimens are from 2001-2002 invertebrate sampling on the Mt Weld altitudinal transect.Th e transect was a satellite project of the International Biodiversity Observation Year and was established and sampled by Forestry Tasmania and the Tasmanian Department of Primary Industries and Water.At the 1100, 1200 and 1300 m sites, six weather-protected pitfall traps (ca 400 mL capacity, ca 90 mm top diameter, fi lled either with undiluted ethylene glycol or the same plus 5% glycerol) were left open for about four weeks every month in the austral summer.Site details and transect history are given in Grove (2004).
Distribution and habitat.Common in caves in the Junee-Florentine karst northwest of Maydena in south central Tasmania (Fig. 9), to at least 200 m depth.
Remarks.N. infi mus is troglomorphic in lacking pigment, and in having more and longer setae, and more slender rings, legs and antennae than its surface-dwelling congener N. summus.A troglobitic polydesmidan from the Cape Range in Western Australia, Stygiochiropus communis Humphreys & Shear, 1993 (Paradoxosomatidae), has been shown to vary considerably from cave to cave in details of gonopod form (Humphreys and Shear 1993).In contrast, there is almost no gonopod variation from cave to cave in N. infi mus.However, populations vary across the Junee-Florentine karst in the degree to which the trunk and legs are attenuated.Growling Swallet cave has the most troglomorphic specimens, with midbody ring length ca 1.3 X prozonite width.In both males and females from Growling Swallet the lower edge of the labrum is extended ventrolaterally on either side as a rounded tab of unknown function (Fig. 4E).
N. infi mus is the only troglobitic millipede so far known from the Junee-Florentine karst.Th e surface-dwelling polydesmidans Paredrodesmus bicalcar, Tasmanodesmus hardyi Chamberlin, 1920 and an undescribed Tasmaniosoma species have also been found in Junee-Florentine caves, but only close to the surface (QVM specimen records).
Although the forests of the Florentine Valley have been repeatedly sampled for millipedes, no surface-dwelling Noteremus species has so far been found there.Th e disjunction between N. infi mus and N. summus is 30-40 km.
Th ere are now two Australian groups fl oating within Dalodesmidea: the Noteremus-Paredrodesmus-Procophorella group discussed above, and the Asphalidesmus group, which is likely to include Agathodesmus steeli Silvestri, 1910 from New South Wales.More needs to be learned about the large and still largely undescribed dalodesmidean fauna of Australia and of New Zealand (Johns 1970) before a satisfactory hypothesis of relationships within the suborder can be proposed.Diagnosis.Diff ers from A. leae and A. parvus in having long, tapering rather than short, convex paranota, and in ozopore opening well away from base of paranotum (at about midheight on body in lateral view) rather than just above paranotal base.

Asphalidesmus golovatchi
Description.Juveniles and cave-dwelling adults unpigmented, but as in Asphalidesmus leae and A. parvus, tergites of surface-dwelling adults are partly encrusted with soil particles and stained light yellow-brown.Surface-dwelling males ca 7 mm long, maximum diameter 0.7 mm, maximum width across paranota 1.2 mm.
Male (Fig. 1C) with head and last ring strongly fl exed to face substrate.Head sparsely setose, as wide as collum; antennal sockets strongly impressed ventrolaterally, separated by about 1.5X socket diameter.Antenna (Fig. 2C) short and thick; antennomere 6 widest and longest; relative antennomere lengths 6>>3>2>(4,5).Collum with anterior edge nearly straight, posterior edge broadly convex, corners blunt.Overall ring widths diminishing gradually from ring 3 posteriorly.Collum, metatergites and paranota with transverse zone of small tubercles (Figs 7A, 7B), each bearing a stout, pointed seta; metazonites also with much smaller, non-setiferous tubercles anteriorly and posteriorly; prozonites with narrow band of longitudinal ridges just anterior to suture, elsewhere uniformly covered with very small protuberances with blunt, rounded tips directed slightly posteriorly.Limbus composed of long tabs with multi-toothed tips and a narrow, outwardly curving medial section.Tergites of rings 2-16 with paired, paramedian, dorsal projections (Fig. 1D), each projection thick, rounded and directed slightly posteriorly; tergites of rings 17 and 18 each with one mid-dorsal, thick, rounded and posteriorly directed projection (dorsal projections absent in juveniles).Paranota of ring 2 greatly expanded laterally and anteriorly and strongly depressed, in lateral view masking head, antennae and collum.Paranota of rings 3-18 set low on body (Fig. 1D), more or less hastate, strongly depressed; lateral extent diminishing gradually from anterior to posterior; anterior and posterior margins scalloped with small, usually discrete, rounded tabs.Ring 3 paranotum slightly overlapping ring 2 paranotum.Pore formula 5, 7, 9, 10, 12, 13, 15-18; ozopore very small, opening in slight depression in low, cylindrical structure well above base of paranotum (Fig. 7B).Sternites somewhat longer than wide, longitudinal and transverse impressions well-defi ned.Pre-anal ring with a few dorsal marginal setae, epiproct not developed, hypoproct paraboloid.Spinnerets in square array (Fig. 8B) in slight depression with low partition wall between dorsal and ventral pairs; sheath separated from seta by annular gap.Anterior legs (Fig. 2C) short; prefemur and femur a little expanded dorsally; relative podomere lengths tarsus>(prefemur, femur)>(postfemur, tibia).No sphaerotrichomes or brush setae; very small tuberculation of metazonite surface extending onto coxa/trochanter.Spiracles (Fig. 7B) opening on short, wide-rimmed elevations; on diplosegments with anterior spiracle above anterior leg and oriented anterolaterally, and posterior spiracle above and about midway between anterior and posterior legbases.Gonopore small, opening on distomedial projection of leg 2 coxa.Legpair 6, 7 bases well-separated to accommodate retracted gonopods, legpair 4, 5 bases a little less separated.
Gonopod aperture ovoid, about one-third prozonite width; posterolateral margins slightly raised.Gonocoxae small, truncated conical, weakly joined distomedially, with a few setae on basomedial and distolateral surfaces.Cannula prominent, inserting in shallow depression on basal surface.Telopodites separate, reaching to legpair 4 bases when retracted.Telopodite (Figs 6B, 6C, 8A) cylindrical, slightly tapering distally, with two slender branches arising at slight constriction at about three-quarters telopodite height; anteromedial branch bending laterally and terminating in 'fi shtail' fork at about two-thirds height of posterolateral branch; posterolateral branch (= solenomere) curving posterolaterally, then slightly anteromedially before fl attening and curling apically with two upright fi nger-like processes arising from curled, fl at tip: a narrower, shorter, more lateral process, and a stouter, taller, more medial process bearing the end of the prostatic groove.Telopodite sparsely setose posterolaterally from near base to distal constriction.Prostatic groove running slightly anteriorly from insertion before running distally to telopodite constriction, then following curve of solenomere to tip.
Female as large as male; posterior margin of epigynum slightly raised medially; cyphopods not examined.
Distribution and habitat.Known from wet eucalypt forest and caves over ca 600 km 2 in far southern Tasmania from ca 100 m to 900 m (Fig. 9).Uncommon in forest.Sympatric over the whole of its range with A. parvus, which also occurs in caves.
Etymology.Adjective, genitive singular, for Sergei I. Golovatch, Russian diplodologist, who has generously given me advice on diplopodological problems and who has taken a particular interest in Asphalidesmus.
Remarks.Cave specimens of A. golovatchi are a little smaller than surface-dwelling specimens in the same stadium.In some of the mature cave specimens the paramedian dorsal projections are greatly reduced, although the median dorsal projections on the last two leg-bearing rings can be clearly seen.
Th e pair in copula from Spring Cave (QVM 23:12971) are in the usual position for mating Polydesmida.Although the male's head is fl exed strongly down, there is still a gap between it and the female's head.
Th e spinnerets in A. leae are arranged as in A. golovatchi, while in A. parvus low partition walls divide the depression housing the spinnerets into four separate compartments (Figs 8B-8D).