A mountain of millipedes IV : Species of Prionopetalum Attems , 1909 , from the Udzungwa Mountains , Tanzania

Two species of the genus Prionopetalum Attems, 1909, are recorded from the Udzungwa Mountains: P. asperginis sp. nov. and P. kraepelini (Attems, 1896). Prionopetalum stuhlmanni Attems, 1914, is synonymized under P. kraepelini. Odontopyge fasciata Attems, 1896, is transferred from Prionopetalum to Aquattuor Frederiksen, 2013, and new illustrations are given. A new illustrated key to species of Prionopetalum is provided.


Introduction
This is the fourth in a series of articles about the millipedes, especially the endemic Afrotropical family Odontopygidae, of the Udzungwa Mountains, Tanzania.For general information on the Odontopygidae and the Udzungwa Mountains see the first article in the series (Enghoff 2014; see also Enghoff &Frederiksen 2015 andEnghoff 2016).
Unlike many other genera of Odontopygidae, Prionopetalum Attems, 1909, is well-defined and easily recognized since the monumental work of Kraus (1960).It is also one of the few odontopygid genera which have been subject of a subsequent comprehensive review (VandenSpiegel & Pierrard 2009).In both of these treatments, a key to species is included, in German and French, respectively.Considering that the vast majority of species of Prionopetalum live in East Africa, where English is much more widely understood than German and French, a key in English is provided here.Two species of Prionopetalum, one of them new, have been collected in the Udzungwa Mountains and are (re)described here.Prionopetalum stuhlmanni Attems, 1914, is shown to be the same species as P. kraepelini (Attems, 1896).Furthermore, Odontopyge fasciata Attems, 1896, which had been included, with some doubts, in Prionopetalum by previous authors, is shown to belong in the genus Aquattuor Frederiksen, 2013.With these adjustments, Prionopetalum now contains 23 described species.

Material and methods
The bulk of the material for this article comes from the zoological collections of the Natural History Museum of Denmark, University of Copenhagen (ZMUC).These specimens were collected during field trips by ZMUC staff and students.Additional specimens from other institutions (see below) were also examined.All specimens are kept in 70% alcohol.Specimens were examined in alcohol under a stereo microscope.Specimens for scanning electron microscopy (SEM) were transferred to 96% ethanol, then to acetone, air-dried, mounted on aluminium stubs or on pieces of flexible aluminium tape and in turn mounted on stubs, coated with platinumpalladium and studied in a JEOL JSM-6335F scanning electron microscope.
As in previous articles in this series, only adult males are considered.A total of 22 adult males of the two Udzungwa species were examined.Figure 1 shows the Udzungwa localities where Prionopetalum specimens were collected.
See Enghoff (2014) for the description standards used.
Unlike many other genera of Odontopygidae, Prionopetalum is well-defined and quite homogeneous, not only in non-sexual characters, but also in gonopod structure.

Diagnosis
(Modified after Kraus 1960 andVandenSpiegel &Pierrard 2009, excluding some non-gonopodal characters which were mentioned by these authors but which are of no diagnostic value.) Prionopetalini in which the anal valves have a raised rim, a well-developed dorsal spine and sometimes a smaller ventral one; limbus with simple, pointed denticles (true of all species after removal of "P." fasciatum, see below).Male legs with ventral soft pads on postfemur and tibia, except on anteriormost and posteriormost legs (true of all species after removal of "P." fasciatum, see below).

Body size
Published body diameters for adult males range from 2.5 mm (P.exaratum, P. glomeratum) to 6.8 mm (P.bifidum), but a male of P. frundsbergi in the ZMUC collection has a diameter of 7.6 mm.Published numbers of podous rings range from 57 (P.exaratum) to 71 (P.bifidum), but the above-mentioned male of P. frundsbergi has 72 podous rings.Table 1 and Fig. 2 summarise the size information and give detailed information for the two Udzungwa species.One has been subtracted from published "segment" numbers because these have traditionally included the telson.Two records have been omitted: "nearly fourty-three" for P. clarum (Chamberlin, 1927) and 93 for P. tanganjikum Verhoeff, 1941 -the latter number is probably a lapsus calami.Based on original measurements and data from the literature (see Table 1).For "other spp." the entries are median values of the intervals in Table 1.The report of "93 (1) Rumpfringen" (corresponding to 92 podous rings) for P. tanganjikum by Verhoeff (1941) has been omitted, as has Chamberlin's (1927) "nearly fourtythree" segments for P. clarum; both are regarded as quite unlikely and are probably erroneous.

Diagnosis
Differs from congeners by the combination of a laterally smooth gonopod coxa, a pointed apical mesad metaplical process subtended by a rounded mesad lobe, a simple, fingerlike proximal telomere process and a simple distal telomere process without secondary branches.
mAle leGs.Postfemora and tibiae with large, soft pads, except on first four to five and several posteriormost leg pairs.Gonopod CoxA (Fig. 3D-F).Lateral margin almost straight, entirely smooth.Mesal margin of proplica straight, proplical lobe (prl) in anterior view almost hidden behind apical expansion of metaplica.Basal part of metaplica with large longitudinal mesad flange (mlf), separated by a deep sinus from an obliquehorizontal, sub-semicircular mesad flange (mof), apical part of metaplica expanded anteriad, forming a pointed process (amp) covering proplical lobe, a rounded mesad lobe (mml) and a slender disto-mesad process (mmp).

Distribution and habitat
Known only from the southern part of the Udzungwa Mts, Udzungwa Scarp Forest Reserve and the Kihansi area.The latter area has become famous because of the Kihansi spray toad, Nectophrynoides asperginis Poynton, Howell, Clarke & Lovett, 1999, which occurred in the Kihansi area but is now regarded as extinct in the wild although a reintroduction programme was started in 2013 (IUCN SSC Amphibian Specialist Group 2015).Zilihona et al. (1998) described in detail the area where I. Zilihona collected her specimens.Altitudinal range 550-750 m asl (cf.Zilihona et al. 1998).

Coexisting species
The sample collected by I. Zilihona also contains another, much smaller odontopygid which will be described in a forthcoming paper.

Remarks
Fig. 3G shows a transverse section (= break) of the basal part (basomere, ba) of the gonopod telopodite.The picture clearly shows how the canal (eg) that continues to the tip of the solenomere is formed as a groove near the surface.This canal has been supposed to be a sperm canal ("Samenrinne" of Kraus 1966, e.g., his figs 28-29), but considering that nothing is known about its function, the neutral term "efferent groove" is preferable (cf.Enghoff 2011).There is also what looks like an entirely internal canal (ic) which is accompanied by dense bundles of tubuli, probably tracheae (Fig. 3H), similar to those recently described by Reboleira et al. (2015) from the base of the vulva of a cambalid millipede.The extent and function of this canal is unknown.

Diagnosis
Differs from all congeners by the multi-cusped proximal telomere process (Fig. 6F).
Colour.After 1-3 years in alcohol head amber below antennae, blackish above; antennae blackish brown; collum blackish brown with light margins; body rings light brown below ozopores and immediately above, dorsally blackish brown with broad middorsal light spot; telson and legs medium brown.

Coexisting species
At Mang'ula, P. kraepelini was collected together with another odontopygid species which will be described in a forthcoming article.

Remarks
Two of the four species of Odontopyge (re)described by Attems (1896) distinguished themselves by the possession of a peculiar, multi-cusped process on the gonopod telopodite.This process, labelled c on Attems' figures, corresponds to the proximal telomere process (tpp) in the sense of the present paper.The two species in question were O. kraepelini, described as new, and a species which Attems identified as O. pardalis (Gerstäcker).He later (Attems 1914) realised that this was not the real pardalis Gerstäcker and offered the replacement name Prionopetalum stuhlmanni Attems, 1914 (pardalis has subsequently been transferred to the genus Calyptomastix Hoffman & Howell, 2012).
The two species, kraepelini and stuhlmanni, are very similar indeed according to the descriptions and illustrations offered by Attems (1896).The only substantial apparent difference concerns the profile of the gonopod coxa.Attems (1896) provided two gonopod drawings of stuhlmanni (as pardalis), and one of kraepelini.The three drawings are all quite different regarding the gonopod coxa profiles, but this is due to the fact that two of them, the one of kraepelini (Attems' fig. 1) and one of those of stuhlmanni (Attems' fig. 8), are based on gonopods macerated in KOH, whereas his fig.7 (of stuhlmanni) is based on unmacerated gonopods.Attems' fig.7 is fully compatible with the present illustrations (Figs 5-6) of specimens from Mang'ula, and side-by-side comparisons of the body and gonopods of Mang'ula specimens with the holotype of stuhlmanni reveal no differences (see Fig. 2 for agreement in body size).The male syntype of kraepelini is devoid of its gonopods, and these are not retrievable elsewhere.I interpret the apparent differences between macerated gonopods of kraepelini (Attems 1896: fig. 1) and stuhlmanni (Attems 1896: fig.8) as being artificial and possibly due to different durations of the KOH maceration, and I therefore synonymize the two names.
The status of "Prionopetalum" fasciatum (Attems, 1896) Attems (1896) described Odontopyge fasciata from Zanzibar, but later (1914) transferred the species to Prionopetalum without further comment.Brolemann (1920) redescribed the species based on new material from Zanzibar and expressed doubts about the genus-level classification of it, suggesting that it might belong in a separate subgenus of Prionopetalum.Kraus (1960), however, accepted the inclusion of fasciata in Prionopetalum.In the most recent treatment of Prionopetalum VandenSpiegel & Pierrard (2009) discussed fasciata, noticing that it differs from other species of Prionopetalum in several characters: the limbus is almost straight, not serrate; there are no ventral tibial pads on male legs; and the proximal telomeral process (tpp) is a simple, rounded lobe.The latter authors reported further new material from Zanzibar as well as from Bagamoyo District, Vula Mountain, Pongwe, on the Tanzanian mainland.
After having examined several specimens from Zanzibar I can now offer a re-classification of this species: Aquattuor fasciatus (Attems, 1896)  In addition, Dr. Nesrine Akkari has kindly examined syntypes of O. fasciata in NHMW and has placed information and images at my disposal.

Remarks
The single intact male among the examined syntypes has 53 podous rings (no apodous rings in front of the telson) and a diameter of 1.91 mm.A syntype at NHMW has the same number of rings and a diameter of 1.86 mm.The specimens from Kizimkazi are all broken, the males have maximum body diameters of 1.55-1.62mm.The male from the sultan's palace has 51 podous rings (no apodous rings in front of the telson) and a diameter of 1.86 mm.
The limbus (Fig. 7A) consists of large, rectangular, easily detached flaps, characteristic of the genus Aquattuor.
The characteristic limbus was not noted by previous authors.Attems (1896) and Brolemann (1920) did not mention the limbus at all.VandenSpiegel & Pierrard (2009: fig. 1b) illustrated the limbus, but their SEM micrograph only showed an irregularly wavy margin.The large limbus lobes were probably all broken off in the specimens examined by these authors (compare the left part of Fig. 7A with fig 1b in VandenSpiegel & Pierrard 2009).In the syntypes from ZMUH the limbus seems to be completely worn off.However, Dr. Nesrine Akkari has kindly provided images based on the syntypes of O. fasciata in NHMW which clearly show the Aquattuor type of limbus, as well as confirm the great similarity in gonopod structure between this species and A. claudiahempae (Fig. 8).The gonopods of the intact male syntype (now dissected by me) closely agree with the illustrations in Figs 7-8.The extra set of gonopods among the syntypes very much resembles fig.6 in Attems (1896), which is based on macerated gonopods, and they were most probably the model for this drawing.
The syntypes of fasciatus as well as the Zanzibar specimens collected by M. Stoltze are very similar to A. claudiahempae Enghoff & Frederiksen (2015), and it is possible that the latter name should be regarded as a junior synonym of A. fasciatus.However, the body diameter of the specimens from Kizimkazi is at the high end, and partly larger, compared to A. claudiahempae from Mt. Kilimanjaro.The two intact specimens from Zanzibar examined here, viz., the intact syntype and the male from the sultan's palace, are definitely much larger.The latter specimens are more similar in size to A. cf.claudiahempae, recorded from the Morogoro Region, Kilosa District, Rubeho Mts by Enghoff & Frederiksen (2015).Attems (1896) gave the (horizontal) diameter of A. fasciatus as 2 mm.Brolemann (1920) gave 55 podous rings (= "56 segments … 1 segment apode") and a diameter of 1.70 mm for an adult male.Whereas "2 mm" is much more than any other specimen in this complex, Brolemann's values would fit A. claudiahempae by extrapolation.
The situation is obviously complicated, and for the time being, A. fasciatus and A. claudiahempae are kept as separate species.

A key to the species of Prionopetalum
This key builds extensively on previous keys provided by Kraus (1960) andVandenSpiegel &Pierrard (2009).The new species described here is included, but Aquattuor fasciatus is excluded, and P. stuhlmanni is treated as a synonym of P. kraepelini (cf.above).I follow VandenSpiegel & Pierrard (2009) in not considering Spinotarsus werneri Attems, 1910 as a species of Prionopetalum, although Kraus (1960) suggested such a relationship (but still listed S. werneri under "species incertae sedis").

Fig. 1 .
Fig. 1.Map of the Udzungwa Mountains, showing the collecting localities for Prionopetalum asperginis sp.nov.(yellow dot) at the southern extremity of the Udzungwa Scarp Forest Reserve and for P. kraepelini (Attems, 1896) (red diamond) at the eastern edge of the Mwanihana Forest Reserve.Based on fig. 1 in Marshall et al. (2010).

Fig. 2 .
Fig. 2. Body size (body diameter / number of podous rings) of ♂♂ of Prionopetalum spp.Based on original measurements and data from the literature (see Table1).For "other spp." the entries are median values of the intervals in Table1.The report of "93 (1) Rumpfringen" (corresponding to 92 podous rings) for P. tanganjikum byVerhoeff (1941) has been omitted, as hasChamberlin's (1927) "nearly fourtythree" segments for P. clarum; both are regarded as quite unlikely and are probably erroneous.

Table 1 .
Numbers of podous rings and body diameter of adult males of species of Prionopetalum.One has been subtracted from published "segment" numbers because these include the telson.None of the species are known to have apodous rings between the last podous ring and the telson.