Cambaloid millipedes of Tasmania, Australia, with remarks on family-level classification and descriptions of two new genera and four new species (Diplopoda, Spirostreptida)

Abstract The Southern Hemisphere cambaloid millipede genera are here assigned or re-assigned to the families Cambalidae Bollman, 1893 and Iulomorphidae Verhoeff, 1924. Tasmanocambala is erected for the three Tasmanian cambalids, T.greenigen. n., sp. n. (type species), T.tasmanicasp. n. and T.taylorisp. n. The new genus is distinguished by a thin, transverse tab at the tip of the anterior gonopod telopodite with a comb of setae immediately behind the tab. The iulomorphid Talomiusweldensisgen. n., sp. n. is described from a single site in Tasmania’s southern mountain district. The new species is unusual among the Australian Iulomorphidae in having a fully-developed, ambulatory leg 1 in the male, and small, compact gonopods.


Introduction
The classification of the cambaloid spirostreptidans has long been unsettled: "Suborder Cambalidea. The systematics of this group are, and will for a long time remain, in a state of particular confusion..." (Hoffman 1980: 82). "Throughout this century great confusion has reigned regarding the classification of the 'Cambalida'..." (Jeekel 1985: 101) "Having had very little experience with cambalidans, I am not in a position to resolve this chaos..." (Shelley 2003: 191). "Confusion still reigns in the various classification databases available on the WWW, and the position of all the "cambalidan" families are chaotic..." (Korsós and Johns 2009: 2).
Especially confusing has been the systematics of the Australasian cambaloid genera, which have even been placed in different orders (Mauriès 1987). In his classification of the Diplopoda, Hoffman (1980) placed three of the Australasian cambaloid genera in Cambalidae Bollman, 1893 and the rest in Iulomorphidae Verhoeff, 1924, which he described as "an assemblage of incongruities" (Hoffman 1980: 84). The 10 Southern Hemisphere iulomorphid genera were "provisionally referred here pending revisionary studies" (Hoffman 1982: 699). Jeekel (1985) accepted Hoffman's grouping as a working classification, and drew a tentative distinction between Cambalidae and Iulomorphidae. In both families the anterior gonopods are thought to deliver sperm, but the posterior gonopods are better developed in Cambalidae. Twenty years later, after examining the eastern Australian cambaloids he had collected in 1980, Jeekel retained the two-family classification and offered another diagnostic character: "The main character used for distinguishing the Cambalidae from the Iulomorphidae is the presence of a well developed flagellum in the anterior gonopods. It remains to be seen if this character is of sufficient importance to separate the two groups. In the family Julidae genera with and without a flagellum may be quite closely related." (Jeekel 2006: 65) Korsós and Read (2012) decided that presence/absence of a flagellum was not of sufficient importance, and placed all Southern Hemisphere cambaloid genera in Iulomorphidae: "...we prefer here to consider all the following 15 genera in the family Iulomorphidae. The only diagnostic difference for separating Iulomorphidae from Cambalidae was the absence of a flagellum in the anterior gonopods (Jeekel 2006a) but this can be a homoplastic, convergent character state as has been shown for the family Julidae (Enghoff 1981, Read 1990. Given this situation, we think it is plausible to combine the very similar Southern Hemisphere (i.e. Gondwanan) elements into one family." (Korsós and Read 2012: 44) Since "Gondwanan" and "Southern Hemisphere (-dwelling)" are not synapomorphies, this action left Iulomorphidae sensu Korsós & Read, 2012 not diagnosable on  morphology. I accepted Iulomorphidae sensu Korsós & Read, 2012 for Amastigogonus Brölemann, 1913, Atelomastix Attems, 1911and Equestrigonus Mesibov, 2017(Mesibov 2017a. However, like Jeekel I have examined two distinct groups of Australian cambaloids: flagella-bearing taxa with relatively well-developed posterior gonopods and non-flagella-bearing taxa with greatly reduced posterior gonopods. The two groups differ ecologically as well, at least in southeast mainland Australia and Tasmania. The first group is most abundant in leaf litter and is common in dry forest and woodland, while the second is mainly associated with rotting wood in higher rainfall areas.
The higher classification of Australasian cambaloids may be resolved in future with the inference of a molecular phylogeny of the group. In the meantime, although I do not wish to further confuse the already muddled classification of the cambaloids, I would like to formalise the distinction between the two Australasian cambaloid groups. I am therefore restoring the Southern Hemisphere genera to the families in which they were placed by Hoffman (1980), with the addition of new Australian genera as assigned by Jeekel (2006) and Mesibov (2017a, and with species numbers from MilliBase (http://www.millibase.org/; accessed 2 January 2019) amended following Korsós and Read (2012) and Mesibov (2017aMesibov ( , 2017b as follows: Cambalidae as recognised here and by Hoffman (1980Hoffman ( , 1982 and Jeekel (2006Jeekel ( , 2009) is a temporary, place-holding taxon and may not be a natural group. It can be distinguished from Iulomorphidae within the Australasian cambaloids by the much greater development of the posterior gonopods and by the presence of a long, very slender flagellum arising medially near the base of the anterior gonopod coxa.
In this paper I describe a new genus and three new species of Cambalidae from Tasmania. I also describe an interesting new species of Tasmanian Iulomorphidae and erect for it a new genus.

Materials and methods
All specimens are preserved in 80% ethanol in their respective repositories.
Photomicrographs were taken with a Canon EOS 1000D digital SLR camera mounted on a Nikon SMZ800 binocular dissecting microscope equipped with a beam splitter. Measurements were made to the nearest 0.1 mm with the same microscope using an eyepiece grid and a reference scale. Photomicrographs used in the figures are focus-stacked composites prepared with Zerene Stacker 1.04. Scanning electron microscope images were acquired digitally using an Hitachi SU-70; specimens were examined after air-drying and sputter-coating with a minimal layer of platinum, then removed from stubs and returned to alcohol. The gonopods of the iulomorphid holotype were temporarily mounted in 1:1 glycerine:water and imaged using an eyepiece video camera mounted on an Amscope binocular microscope. Preliminary drawings of the gonopods were traced from printed copies of images, and drawings were then edited by reference to the actual specimens. Images and drawings were prepared for publication using GIMP 2.8.
Maps were drawn with QGIS 2.4. Latitude/longitude figures in the text (all based on the WGS84 datum) are given in decimal degrees to four decimal places, together with a spatial uncertainty. In some cases, collecting site locations have been upgraded from UTM grid references (on original labels, with the AGD66 datum), based on advice from collectors and the latest digital mapping of Tasmania. The spatial uncertainty figure covers the likely error in the location upgrade as well as my estimate of the likely error in the original location. All specimen records referred to in the text are in Supplement 1.
I follow Enghoff et al. (1993) in counting trunk rings by excluding the telson and giving podous + apodous ring counts, e.g. "(55+1) rings", and I give the upper limits of the count ranges I observed rather than count frequencies. Abbreviations: QVMAG Queen Victoria Museum and Art Gallery, Launceston, Australia TMAG Tasmanian Museum and Art Gallery, Hobart, Australia ZMUC Zoological Museum, Natural History Museum of Denmark, Copenhagen, Denmark
Diagnosis. Differs from the other seven Southern Hemisphere cambalid genera by the anterior gonopod telopodite ending in a thin tab just anterior to an apical, transverse comb of setae; from Dimerogonus and Eumastigogonus in having a smoothly rounded apex of the coxal process, without a medial extension; from Proscelomerion in lacking a pseudoflagellum and in having a rounded rather than acuminate tip on the coxal process; from Apocoptogonus and Euryischiogonus by the flagellum not having a bifurcate tip; from Stenischiogonus by the lack of a distinct medial lobe on the tip of the anterior gonopod telopodite; and from Zinagon by the male leg 1 femur being much wider than the more distal three podomeres and by the anterior section of the posterior gonopod lacking a needle-like extension.
Aperture on ring 7 (Fig. 2D-F) cordate (apex to rear), the lateral margins slightly raised. Anterior and posterior gonopods forming small, compact structure, tilted posteriorly in ring 7. Coxa of anterior gonopod (Figs 1D, E; 2) about as long as telopodite or a little shorter, anteroposteriorly flattened and with large posterior concavity holding telopodite; apex rounded and very thin, directed distomedially; flagellum not bifurcate, arising medially on coxal base and curving first posterobasally, then distally, then anteriorly. Telopodite of anterior gonopod (Fig. 1D, E) not as wide as coxa, tapering distally, posterolaterally slightly excavate with a few very short setae in deepest portion of excavation near base ("rudimentary terminal podomere" of Korsós and Read (2012)); telopodite ending in translucent, rounded tab with comb of setae just posterior to tab, the setae shorter than tab. Posterior gonopods (Fig. 1F) reaching ca 2/3 height of anterior gonopods; anteriorly divided by deep, oblique groove into anterolateral and posteromedial sections, subequal in height; anterolateral section tapering at ca 1/2 section height from wide base to apically rounded lamina; posteromedial section with apex stout, bluntly rounded and tipped with sparse brush of short setae, and with row of very short setae along anteromedial margin of section.
Females like males in overall appearance but noticeably stouter; vulvae not examined. Distribution. So far known only from Tasmania, Australia. Remarks. Males of Tasmanocambala gen. n. species are identifiable by examination of the tip of the anterior gonopod, even with the gonopods lying in situ in ring 7: there is a terminal fringe of setae apparent behind a thin, translucent, anterior tab. The type species is likely to be a species complex (see Remarks on the type species) and the taxonomy of this genus would greatly benefit from genetic analysis.
The deep medial groove on the male leg 1 tarsus (Fig. 1C) is hard to detect with optical microscopy. It may be an autapomorphy for the genus, or it may be present in other cambaloid millipedes but overlooked by describers.  Other material. 89 males and 76 females from 44 sites other than the type locality, in QVMAG, TMAG and ZMUC; see Supplement 1 for details.
Distribution. Widespread in Tasmania but not yet recorded from the Midlands or the Northeast (Fig. 5A).
Name. In honour of Robert "Bob" Green (1925Green ( -2013, Tasmanian zoologist and former Curator of Zoology at QVMAG. Green collected the type specimens of T. greeni sp. n. during his 15-year study of the impact of logging operations at Maggs Mountain in northwest Tasmania.
Remarks. This species is likely to be a species complex, as there are geographically correlated variations in body size and in the length, position and distinctiveness of the horizontal striae on the trunk rings. However, I have not been able to observe any consistent, corresponding differences in gonopod structure. I chose the type specimens from the largest form in the putative species complex; this larger form mainly occurs in mid -to high-elevation areas in northwest and central Tasmania. Other material. 53 males, 47 females from 42 sites other than the type locality, in QVMAG; see Supplement 1 for details.
Name. This species appears to be endemic to Tasmania. Diagnosis. Distinguished from T. greeni n. sp. and T. taylori n. sp. by the annular striae on the prozonite and by the larger number of setae in the apical comb on the anterior gonopod telopodite.
Description. As for the genus, with the following details: in life, legs distinctly red-coloured in life (colour fades in alcohol) and strongly contrasting lightcoloured annular band at rear of metazonite; male/female to 52+1/56+1 rings, 2.5/2.9 mm in midbody diameter. Trunk rings with variable number of annular striae on prozonite anterior to suture, most obvious ventrally (Fig. 2B). Anterior gonopod coxa (Figs 1E, 2E) with broad tip, usually wider than in T. greeni n. sp., usually shorter than telopodite; setal comb behind apical tab extending full width of telopodite tip and with 7-10 setae; tab sometimes with small finger-like projection laterally.
Distribution. East of Tyler's Line (Mesibov 1994) in the Northwest, but apparently absent from the northeast of the main island and from islands in Bass Strait (Fig.  5B). The polydesmidan Tasmanodesmus hardyi Chamberlin, 1920 is similarly distributed (Mesibov 2004).
Remarks. Tasmanodesmus tasmanica n. sp. is the largest cambalid species in Tasmania and could be confused at first glance with the similar-sized iulomorphid Equestrigonus tasmaniensis Mesibov, 2017. The two spirostreptidans have not yet been found to cooccur, but their ranges may overlap in wet forest south of Wynyard and near Blessington.
Tasmanodesmus tasmanica sp. n. is very abundant in the Dial Range south of Penguin (i.e., around the type locality), where it can readily be found in and under damp leaf litter in wet eucalypt forest at any time of year.  Paratypes. In QVMAG: 2 females, details as for holotype, QVM:2017:23:0041. Other material. 18 males, 41 females and 3 juveniles from 11 other sites on Flinders Island and 4 sites on Prime Seal Island, in QVMAG and TMAG; see Supplement 1 for details.
Name. In honour of Robert Taylor, collector of the type specimens. During his 13 years in Tasmania, Taylor instigated and managed a number of forest conservation projects that employed the author and other local zoologists as field workers and specimen processors. Material collected for those projects has been a valuable taxonomic resource for the author and others, and will continue to be valuable for years to come.
Diagnosis. Distinguished from T. tasmanica n. sp. by the absence of annular prozonite striae and from T. greeni n. sp. by the strong mediad curvature and greater narrowing of the tip of the anterior gonopod coxa.
Distribution. So far known from Flinders and Prime Seal Islands at the eastern end of Bass Strait between Tasmania and Victoria (Fig. 5A).
Remarks. Tasmanodesmus taylori sp. n. is not greatly different from central Tasmanian forms of T. greeni n. sp., and the narrowing of the tip of the anterior gonopod coxa is variable from specimen to specimen.
Diagnosis. Distinguished from all other genera of Iulomorphidae by the male first legs having a reduced prefemur, but with the four more distal podomeres appearing as in normal walking legs, including a normal claw on the tarsus; and distinguished from the other nine Australian iulomorphid genera by the small size of the gonopods relative to ring 7 diameter, by the strong medial curvature of the coxal process on the anterior gonopod, and by the bare, posteriorly curving apex of the anterior gonopod telopodite terminating in the opening of the prostatic groove.
Distribution. Known only from the type locality (Fig. 5B). Remarks. When sorting spirostreptidan millipedes for an article on Tasmanian Iulomorphidae (Mesibov 2017a), I set aside the three Mt Weld males as "Cambalidae", because the males had a small, compact gonopod complex like the Tasmanian cambalids described above, and the legs lacked the prefemoral tabs found in Australian Iulomorphidae. The males also had apparently ambulatory first legs, which so far as I am aware have not been reported before in any iulomorphids. When preparing the current article, I removed the gonopods of two of these "cambalids" and cleared and imaged one of the undissected complexes. Unfortunately, I then lost the two gonopod complexes, leaving only one of the three males intact. Rings 7 and 8 of that male were removed for SEM imaging of the gonopod complex (Fig. 3D), but with only a very thin coat of metal applied. The rings were returned to alcohol and the gonopods dissected and illustrated here; this specimen has been designated the holotype.
I regret not having additional material of T. weldensis n. sp. for study and description, but the type locality is in Tasmania's southern mountain district, which in 2019 remains a remote and little-sampled wilderness area. The three known specimens of T. weldensis n. sp. were in pitfall traps emptied on 28 March 2012 at 600 m on Mt Weld, during a biological monitoring study along an altitudinal transect. They were among ca 50 Amastigogonus verreauxii (Gervais, 1847) (Iulomorphidae) in pitfalls emptied on the same day at the same elevation (A. verreauxii records in Mesibov (2006Mesibov ( -2019). The Mt Weld study generated its invertebrate samples in 2001-2002 and again in 2011-2012. I did not observe any other T. weldensis n. sp. specimens among the millipedes pitfall-trapped in the two sampling periods.

Discussion
There are far fewer specimens of Cambalidae in Tasmanian museum collections than of Iulomorphidae. My field experience over 45 years in Tasmania is that this difference is not due to sampling bias, but reflects the patchy distribution of Cambalidae. Where they occur, however, Cambalidae are often abundant. This "abundant but patchy" distribution is also characteristic of procyliosomatid Sphaerotheriida in Tasmania. Hundreds of pill millipedes can be found in small patches of richly organic forest soil in some areas, while none are seen in apparently identical macro -and microhabitats nearby.
The patchiness of cambalids in Tasmania accords with what Jeekel (2006) observed on the Australian mainland: "During a collecting trip through New South Wales by the author and his wife between 27.X. and 12.XI.1980 only five sites out of 32 yielded representatives of the family. Since each sample consisted of a different species, it is obvious that Cambalidae are quite local in their occurrence. Possibly also their appearance in the upper layers of the soil is restricted to periods of favourable weather conditions. Under such conditions populations may be quite numerous" (Jeekel 2006, p. 65).
Jeekel found no Cambalidae in Victoria or Tasmania (Jeekel 2009, p. 83), but I have observed specimens of Victorian Cambalidae in Museums Victoria and have also collected cambalids in South Australia (specimens deposited in the South Australian Museum).
The almost complete lack of cambalid records in northeast Tasmania is curious, as the region has been intensively sampled for millipedes over many years and hundreds of iulomorphid Spirostreptida have been collected there (see fig. 7 in Mesibov (2017a)). Two, possibly introduced, cambalid species are known from Cuckoo Plantation, which is a long-established Pinus and Eucalyptus plantation near Scottsdale in northeast Tasmania (QVMAG specimen lots QVM:2017:23:0054 and QVM:2018:23:0003). These two species, one of which appears to be a Dimerogonus, may have been carried to Tasmania from mainland Australia on forestry equipment. I have not found additional specimens of either of these species in the Plantation or elsewhere in northeast Tasmania, despite careful recent searching. A third QVMAG specimen lot (QVM:2017:23:0016) contains a female and a juvenile of a possible cambalid from a logged forest area northeast of Goulds Country in the Northeast.
Talomius n. gen. is the fourth iulomorphid genus to be recognised in Tasmania. Like the two Tasmanian Atelomastix species, T. weldensis n. sp. may have a restricted distribution, but it also possible that the three spirostreptidans will be found to have substantial ranges when Tasmania's southern and southwestern wilderness areas are carefully sampled for millipedes in future.