A mountain of millipedes VI . New records , new species , a new genus and a general discussion of Odontopygidae from the Udzungwa Mts , Tanzania ( Diplopoda , Spirostreptida , Odontopygidae )

Damacornu gen. nov. (type species: D. transversum gen. et sp. nov.), Geotypodon papei sp. nov. and Spinotarsus fortehamatus sp. nov. are described, and Helicochetus dimidiatus (Peters, 1855), H. mutaba Kraus, 1960 and Hoffmanides dissutus (Hoffman, 1963) are recorded from the Udzungwa Mts, Tanzania. A complete overview of the 39 odontopygid species now known from the Udzungwa Mts is given, including notes on endemism, biogeographical relationships and altitudinal distribution patterns.


Introduction
The very rich fauna of odontopygid millipedes from the Udzungwa Mts has been the subject of five recent papers (Enghoff 2014(Enghoff , 2016a(Enghoff , 2016b(Enghoff , 2016c;;Enghoff & Frederiksen 2015).The present contribution deals with the remaining Udzungwa odontopygids in the very large collections of Tanzanian millipedes in the Natural History Museum of Denmark and the Virginia Museum of Natural History (but see note on the genus Aquattuor Frederiksen, 2013).Three new species are described, one of them belonging to a new genus, and three previously described species are recorded for the first time from the Udzungwa Mountains.The collecting sites for the treated species are shown in Fig. 1.A list of the known odontopygid fauna of the Udzungwa Mountains, now comprising 39 species, is included, as well as notes on distribution and endemicity.

Material and methods
The material described in this paper derives from the zoological collections of the Natural History Museum of Denmark, University of Copenhagen (ZMUC) and the Virginia Museum of Natural History (VMNH).Methods of study are as in the previous papers in this series, as detailed by Enghoff (2014).
A considerable new collection of Aquattuor spp.from the Udzungwa Mts has recently been acquired by the ZMUC, the treatment of which awaits a renewed study of the genus by Sara Frederiksen.

Other included species
None.

Diagnosis
A genus of Odontopygidae-Prionopetalini characterized by the following combination of characters: gonopod coxa with a mesal spineless metaplical shelf, without a basad metaplical spine; telopodital torsotope with a long spine; solenomere whip-like, with a basal spine but otherwise without outgrowths, apically fluted and spiraled; telomere a longitudinally folded lamella; distal half of telomere bent at right angles in relation to basal half, with a row of teeth on internal surface.

Remarks
In the key to genera of "Odontopyginae" by Kraus (1966), Damacornu transversum gen.et sp.nov.keys out to the genus Odontopyge, and in the key to species of "Odontopyge" by Kraus (1960) it keys out in couplet 11 ("O." citernii Silvestri, 1898 and"O." difficilis Silvestri, 1895).The new species is, however, clearly different from these two species in which there is no metaplical shelf and the proximal solenomeral spine (ps) ("Tibialdorn") is rudimentary.None of the "orphaned" species assigned to Odontopyge by earlier authors (see Enghoff 2016a) can be adopted by Damacornu gen.nov., which for the time being remains monotypic.

Etymology
The name is a Latin adjective and refers to the transverse yellow markings on the body rings.
Colour.After 35 years in alcohol blackish, anterior part of head and posterior part of metazona amber; a transverse yellow band on prozona extending from midline ca.halfway down to ozopores; legs yellow.
Collum.With a marginal furrow, followed by two abbreviated and one complete submarginal furrow.Body ringS.Almost perfect cylinders, not vaulted; suture straight; ozopores starting from ring 6, placed ca ⅓ of metazona length behind suture.18-22 metazonital striae reaching up to ca one metazonite length below ozopore.
anal valveS.With a well-developed dorsal tooth and a small ventral one; margin barely raised, setiferous tubercles inconspicuous.limBuS (Fig. 2A).Divided into simple, pointed teeth, each tooth > 3 times longer than broad.male legS.With adhesive pads on postfemur and tibia.

Female
Unknown.

Diagnosis
(From Enghoff 2016a, slightly modified.)A genus of Odontopygidae-Prionopetalini characterized by: a long basad metaplical spine on the anterior side of the coxa; a compact torsotope; lack of pretorsal or torsal spines/processes; a pronounced post-torsal narrowing without spines; a division of the telopodite into solenomere and telomere immediately distal to post-torsal narrowing; a spine emerging from the base of the telomere; a slender, whip-like, smooth solenomere without any outgrowths or appendages (except sometimes a tiny subapical spine); a telomere consisting of various lobes and lamellae with largely smooth margins.
Geotypodon papei sp.nov.urn:lsid:zoobank.org:act:3D36FD3A-B534-422E-8354-D0EDCAB2A748Fig. 3 Diagnosis Differs from all other species of Geotypodon, except.G. carli (Kraus, 1960), by the combination of a very long, slender metaplical spine and a relatively simple telomere mainly consisting of a longitudinally folded sheet.Differs from G. carli by the shape of the basad, metaplical, spine-like process (spine strongly expanded in basal part in G. papei sp.nov., not so in G. carli).See also Table 1.

Etymology
The
Colour.After 3 years in alcohol, head reddish brown to dark brown, antennae dark brown, collum, body rings and telson dark grey without contrasting markings, posterior ¼ of metazona amber, legs reddish brown.
Collum.With a marginal and two submarginal furrows.
Body ringS.Almost perfect cylinders, not vaulted; suture straight; ozopores starting from ring 6, placed midway between suture and limbus.13-15 metazonital striae, reaching up to ca one metazonite length below ozopore.
anal valveS.Each with a stout dorsal spine; ventrally slightly angular; margins raised; each valve with three marginal setae not borne on tubercles.limBuS (Fig. 3C).Consisting of isolated hand-like lobes with a length of ca 10 μm, each with 5-9 'fingers' and each with two ridges running along its length.male legS.With tiny postfemoral ventral pads from leg-pair 3-5, from leg-pair 6 postfemoral pads large, covering entire podomere, but gradually smaller posteriorly before disappearing completely shortly before posterior end.Tibiae with tiny ventral pads from leg-pair 3 until somewhere between gonopods and mid-body.

Female
Unknown.

Remarks
In the genus key of Kraus (1966), Geotypodon papei sp.nov.easily keys out in the last couplet (33), where the choice is between Patinatius Attems, 1928 and Odontopyge Brandt, 1841.In the couplet, the two genera are distinguished by Patinatius having the "Lateralblatt" prolonged into a cone or extended into a long spine vs Odontopyge not having such modifications.Couplet 33 in Kraus (1966) is copied from couplet 33 in Kraus (1960) with the difference that in the 1960 version, this couplet separates Odontopygista Kraus, 1960 from Odontopyge.Kraus (1966) synonymised Odontopygista with Patinatius.In both versions of couplet 33, Kraus made an error because the part of the gonopod coxa that has a cone-or spine-like outgrowth is the metaplica, which in Kraus' terminology is the "Medialblatt" (as also used in his genus and species descriptions).This aside, one might argue that the long, sigmoid terminal spine (tsp) could be the homologue of such a "long spine", but Geotypodon papei sp.nov.disagrees with the definition of Patinatius in several respects, including the limbus (spatulate and multicusped in Geotypodon papei sp.nov., rounded lobes in Patinatius), the internal surface of the telomere (smooth in Geotypodon papei sp.nov., with a row of denticles in Patinatius) and the tip of the solenomere ("thumb and lamella" in Geotypodon papei sp.nov., simply pointed in Patinatius).Geotypodon papei sp.nov.therefore keys out to Odontopyge in Kraus' keys, and as explained by Enghoff (2016a) this generic name is not available for other than a small group of species very different from Geotypodon papei sp.nov.In the key to species of "Odontopyge" in Kraus (1960), Geotypodon papei sp.nov.smoothly keys out to carli Kraus, 1960, one of the species that was transferred to Geotypodon by Enghoff (2016a), and the two species are indeed very similar.There are, however, enough differences to regard them as separate species, cf.Table 1.

Distribution
Known from three sites in the northeastern part of the Udzungwa Mts, at moderate altitudes (339-650 m a.s.l.).

Remarks
The material from the Udzungwa Mts includes specimens of two species of Helichochetus.Separate descriptive notes are given for each species, but two characters shared by both species may prove to be genus-characteristic and are therefore commented on here.limBuS (Fig. 4A, C, E).Kraus (1960) characterized the limbus of Helichochetus (translated from German): "Limbus divided by rounded indentations into processes which each carry three to six nail-like points.Limbus very broad, ca. 10 × as broad as the processes, with a silky sheen due to the characteristic surface sculpture (exception H. laciniatus)".The limbus of the two studied species, especially H. mutaba, is indeed very broad (Fig. 4A, C), and the processes and nail-like points agree fully with Kraus' description.
It is not clear from his publications (Kraus 1960(Kraus , 1966) ) what he meant by the "characteristic surface structure".The SEM images (Fig. 4A, C, E) provide no clue.There is, however, another remarkable detail, visible in Fig. 4E: under the limbus proper, there is a "sub-limbus" consisting of rounded, almost semicircular lobes.
CytoSCute dentiCleS (Fig. 4B, D, F).The "cytoscutes", i.e., cuticular "cells" each corresponding to a hypodermis cell (Fusco et al. 2000) very often carry a row of tiny denticles along one edge.In both studied species of Helicochetus, these denticles are much larger than normal and point away from the cuticular surface rather than lying parallel to it.
anal valveS.Each with a moderate dorsal spine; margins raised; each valve with three marginal setae not borne on tubercles.limBuS (Fig. 4A).See Remarks for genus.male legS.With postfemoral ventral pads from first post-gonopodal leg-pair 3-5 until ca midbody; anteriormost and posteriormost pads small, intermediate ones covering entire ventral side of postfemur.

Female
Unknown.

Distribution and habitat
Known from several localities in Mozambique and Tanzania (Enghoff et al. 2016).
Verhoeff (1901) quoted a report on H. dimidiatus (as Odontopye attemsi Verhoeff, 1901) from the small island of Kwale (off the southern Tanzanian coast) where this species appeared in very large numbers, destroyed the islander's crops and forced them to move their agricultural activities to the mainland.

Diagnosis
Resembles H. digititarsus Kraus, 1957, H. gregorii (Pocock, 1896) and H. monodon Kraus, 1960 in the combination of a broad, shovel-like process ("Tibialdorn") at the level of the post-torsal narrowing; absence of ventral pads on male tibiae; presence of a basal lamella ("Grundblatt") on the telomere; tip of telomere not drawn out into long process, with several short to moderately long spine-like outgrowths from the margin.Differs from these species in the shape of the basal lamella: slender, pointed and curved (with a few denticles on the concave margin in the new specimens).

Description Male
Based on studied specimens.Details from original description (Kraus 1960) in parentheses.Both of the new specimens are strongly fragmented and completely faded.
Collum.With two complete furrows on each side, none of them marginal (two distinct furrows, marginal furrow less distinct).

Female
Unknown.

Distribution
Hitherto known only from the type locality, Mutaba in the vicinity of Kirungu (= Baudoinville), in the Democratic Republic of the Congo.The two records from Tanzania, Nyumbenito Mt in the Udzungwa Range and Mazombo N of the Udzungwas, thus represent a considerable range extension.See, however, Remarks below.

Remarks
Helicochetus mutaba belongs to a group of very similar nominal species characterized by the following combination of key characters: a shovel-like lamella ("Tibialdorn") at the level of the post-torsal narrowing; no ventral pads on male tibiae; telomere with a basal lamella ("Grundblatt"); tip of telomere not drawn out into a long process, with several short to moderately long, spine-like outgrowths from the margin.This group consists of H. digititarsus Kraus, 1957 (Democratic Republic of the Congo), H. gregorii (Pocock, 1896) (Kenya), H. monodon (Kraus, 1960) (Zambia) and H. mutaba (Democratic Republic of the Congo, Tanzania).Table 2 shows the differences between these four species.The outline of the basal lamella of the telomere differs between the species, but it is not unlikely that future collections will bridge the small morphological gaps and that some or all of the four species will have to be synonymized.Enghoff et al. (2016) recorded H. digititarsus from two Tanzanian localities: 1) Mara Region, Serengeti District, Seronera, Serengeti National Park; 2) Tabora Region, Nzega District, Nzega City.In light of the above discussion, it cannot be excluded that these specimens are conspecific with the one recorded here as H. mutaba.
The vial containing the male from Mazombo also contained a microvial with a mite which was presumably collected in association with the millipede.The mite was infested with numerous thalli of the fungus Rickia sp.(Ascomycota, Laboulbeniales).Several species of mites are known to be closely associated with millipedes (Farfan & Klompen 2012), and several species of Rickia Cavara are known to use mites as their host (Tavares 1985;Seeman & Nahrung 2000), but this seems to be the first known instance of a Rickia-mite-millipede association.

Remarks
This genus, remarkable for is widely open gonopod coxa, was erected by Kraus for a species originally placed in the large, mainly southern African genus Spinotarsus Attems, 1909.By doing so, Kraus deprived Tanzania of the genus Spinotarsus, an action neutralized by the description below of a new Tanzanian species of Spinotarsus.

Description Male
Based on studied specimens.Details from previous descriptions (Hoffman 1963;Kraus 1966) in parentheses.Considering that two detailed descriptions of H. dissutus are thus available, and that the studied specimens agree completely with both of them, only a few details are highlighted here.

Female
Unknown.

Distribution
Helicochetus dissutus was described from "Morogoro" (Hoffman 1963); this does not necessarily refer to the city of Morogoro, ca 150 km NE of Udzungwa National Park, but possibly to anywhere in the Morogoro Region.Enghoff et al. (2016) also recorded the species from the Morogoro Region, Uluguru Mts, Morningside.

Remarks
The vast majority of species of Spinotarsus occur in southern Africa; only two, viz, S. serrulatus Kraus, 1958 andS. terrestris (Attems, 1935), are from the Democratic Republic of the Congo.Two species described from other parts of the Afrotropical region have subsequently been transferred to other genera: S. caboverdus Pierrard, 1987 (Cape Verde Islands) was transferred to the genus Bandeirenica Kraus, 1960by Hoffman (2000); S. dissutus Hoffman, 1963 (Tanzania), was made type of the genus Hoffmanides by Kraus (1966), see above.The new species described below is thus the first true Spinotarsus from Tanzania.

Diagnosis
A species of Spinotarsus characterized by the very large, regularly curving lateral metaplical spine, in combination with the presence of a spine at the base of the solenomerite and a row of denticles on the terminal part of the telomere.

Etymology
The name is a composite Latin adjective, meaning "with a strong hook" and referring to the strong lateral, hook-shaped spine of the gonopod coxa.

Description Male
Size.Length ca 6 cm, diameter 5.0 mm, 59 podous rings, no apodous rings in front of telson.
Colour.Apparently quite well-preserved after 3 years in alcohol: head below antennae yellowish; head above antennae blackish, parietal and interocular furrows contrastingly yellow; antennae brown, collum blackish with lighter margins; body rings light grey, metazona with black sputtering, posterior ⅓ of metazona amber; a yellow dorsal longitudinal stripe on posterior ¼ of body; legs yellow; preanal ring grey, dorsally yellow; anal valves blackish with lighter margin; subanal scale yellow.
Head.Parietal and interocular furrows present; 7 supralabral setae.Antennae reaching to ring 6 when folded back.Eyes reaching to median tangent to antennal socket, each with nine rows of ommatidia, 7 horizontal rows, and a total of ca 38 ommatidia each.
Collum.Lateral lobes subrectangular, each with two furrows.Body ringS.Unvaulted; metazona and posterior part of prosoma with numerous short very fine longitudinal furrows; ozopores ca ⅓ of metazona length behind straight suture.
anal valveS.Each with a stout dorsal spine and a tiny ventral knob-like spine; mesal margin slightly raised, setiferous tubercles indistinct, not on ravelins.limBuS (Fig. 9G).With long (more than twice as long as broad), pointed lobes.
gonopod Coxa (Fig. 9F).Slightly more than twice as long as broad.Proplica ending in small proplical lobe (ppl).Metaplica longer than proplica, apically forming a "cucullus" (cu).Mesal margin of metaplica with a curved lamellar process (mlp) just distal to exit point of telopodite ("arculus").Metaplica laterally with a very large spine (msp) curving distad in a latero-posterior plane.Proplical lobe, lateral metaplical spine and a stripe parallel to lateral metaplical margin from spine to tip of cucullus black (not visible on SEM image).

Female
Unknown.

Remarks
The basal spinose lamella the telomere (bla) is a key character for Spinotarsus (the name actually refers to the lamella -the telomere was formerly referred to as the tarsus).Many species of Spinotarsus possess an additional apomorphy, namely a darkly sclerotized ridge on the posterior surface of the telomere, but such a ridge is absent in S. fortehamatus sp.nov.In the most recent key to species of Spinotarsus (Kraus 1966: 110-113), the new species keys out without problems to couplet 60, as follows (couplet texts translated from German and adapted to current terminology, characters of S. fortehamatus sp.nov. in bold): couplet 1: Gonopodal metaplica with one or several dark sclerotized lateral spines which stand clearly out in strict oral view → couplet 38: One or more spine(s) ("Postfemoraldorn") at base of solenomere present → couplet 59: One spine at base of solenomere → couplet 60. "Couplet" 60 is, however, deficient in the key, which can be seen by comparison with the older key by Kraus (1960: 122, "couplet" 42).The "couplet" in question is actually a "triplet", but the third option is lacking in the 1966 key.In translation, "triplet" 60 in Kraus (1966)  Continuing in couplet 65, the new species fulfils the second alternative: Lateral metaplical spine directed ± apicad → couplet 68: Metaplica apically without a process or just with a delicate hyaline, spoonshaped appendage → couplet 81: Terminal lamella of telomere with spiny margin and 1-2 longitudinal rows of denticles on the internal surface (→ couplet 82) vs Terminal lamella of telomere completely smooth, internal surface of telomere without obvious spine rows (→ couplet 84).The new species has a smooth margin of the terminal part of the telomere, but it also has a very obvious row of spines on the internal surface of the telomere.It thus fits one criterium of each alternative.In addition to its unique combination of key characters, it also differs from congeners in the shape of the lateral metaplical spine: none of the almost one hundred described species of Spinotarsus have an equally large, equally smoothly curved lateral metaplical spine.Those which come closest in this character, i.e., S. demotus Kraus, 1966 (Angola), S. kruegeri Kraus, 1966 (South Africa), S. lanceolatus Kraus, 1966 (Zimbabwe) and S. pusillus Kraus, 1966 (South Africa), differ clearly in other characters.
The lack of ozopores on one or both sides of certain body rings is a highly unusual trait.In juliformian millipedes, ozopores are normally present in an uninterrupted series usually starting on body ring 6, rarely on ring 5, exceptionally on ring 3, and continuing to include the last podous ring.Up to now, the only exception to this rule is constituted by two species of the family Mongoliulidae, in which ozoporeless body rings occur in a regular pattern along the body (Enghoff et al. 2017).

Discussion
The Odontopygidae of the Udzungwa Mountains Thirty-nine species of Odontopygidae (Table 3) are now known from the Udzungwa Mts as a result of the series of studies entitled "A mountain of millipedes" (Enghoff 2014(Enghoff , 2016a(Enghoff , 2016b(Enghoff , 2016c;;Enghoff & Frederiksen 2015;this paper).Previously, just one species (Chaleponcus dabagaensis Kraus, 1958) was known.Based on the examination of a huge collection of millipedes from the Udzungwas, Odontopygidae is clearly the most species-rich millipede family in these mountains.Paradoxosomatidae are also represented by many species, whereas such families as Spirostreptidae, Oxydesmidae, Gomphodesmidae, Chelodesmidae, several families of "micropolydesmoids" and Stemmiulidae seem to have no more than one or two handfuls of species each.
Out of the 39 odontopygid no less than 35 are known only from the Udzungwa Mts and should be regarded as Udzungwa endemics.No phylogenetic analyses are available for any Udzungwan odontopygid, but nevertheless some preliminary geographical patterns may be inferred (Fig. 10): -The endemic Chaleponcus dabagaensis group seems to be most closely related to C. altirungwensis Enghoff, 2017 from Mt Rungwe, SW of the Udzungwas, see Enghoff (2017aEnghoff ( , 2017b)).3.
-The genus Aquattuor (five endemic species in the Udzungwas) has a further species endemic to the E Usambara Mts, and still one (or two?) species recorded from several sites in NE Tanzania, where they may have arrived by means of human transportation (Enghoff 2016b;Enghoff & Frederiksen 2015).-Hoffmanides dissutus (Hoffman, 1963) is also known from the Uluguru Mts.
-Geotypodon papei sp.nov. is very similar to and presumably closely related to G. carli Kraus, 1960 (Democratic Republic of the Congo).-Helicochetus mutaba is also known from the eastern Democratic Republic of the Congo.
Odontopygid "tracks" thus connect the Udzungwa Mts with other mountain blocks in the (extended) Eastern Arc (Aquattuor, Chaleponcus and Hoffmanides dissutus, but surprisingly also with areas in the Democratic Republic of the Congo further west. The altitudinal distribution of the odontopygids (Fig. 11) shows a striking pattern: the endemic species swarm constituted by the Chaleponcus dabagaensis group (Enghoff 2014) occupies high altitudes (species medians ranging from 1520 (C.circumvallatus Enghoff, 2014) to 2100 m a.s.l.(C.tintin Enghoff, 2014).Other endemic species occur at lower altitudes (medians: 400-1799 m a.s.l.), whereas three out of four non-endemic species all occur at low altitudes (medians: 300-399 m), as might be expected.

Fig. 10 .
Fig. 10.Geographical relationships of the Odontopygidae of the Udzungwa Mts.See main text for explanation.No relationships are shown for the three endemic genera (Casuariverpa Enghoff, 2016, Utiliverpa Enghoff, 2016 and Yia Enghoff, 2016), nor for the Udzungwa species of the large genus Spinotarsus Attems, 1909, nor for the widespread species Helicochetus dimidiatus(Peters, 1855) and Prionopetalum kraepelini(Attems, 1896).Base map by permission of the Eastern Arc Mountains Conservation Endowment Fund.
name honours Thomas Pape, collector of the holotype and a leading figure in the Eastern Arc program of the Natural History Museum of Demark.