Revision of the genus Attemsostreptus Verhoeff, 1941 with description of a new species from Tanzania and notes on the tribe Trachystreptini Cook, 1896 (Spirostreptida, Diplopoda)

We revise the genus Attemsostreptus Verhoeff, 1941 based on type material of the type species, A. costatus Verhoeff, 1941, synonymise A. orobius (Kraus 1958) with A. costatus and describe a second species of the genus, A. refl exus sp. nov., collected from Kimboza Forest Reserve in Tanzania, and discuss the dubious tribe Trachystreptini.


Introduction
The genus Attemsostreptus was fi rst erected by Veroheff in 1941 to accomodate the species he then described from the mountains in Tanzania's Morogoro Region, Attemsostreptus costatus Verhoeff, 1941. The genus remained monotypic for over forty years until Krabbe (1982) assigned to it a second species Attemsostreptus orobius (Kraus, 1958) described in the genus Lophostreptus (Kraus 1958). In her work on the systematics of Spirostreptida, Krabbe (1982) provided a diagnosis for the genus based on the gonopod structures and speculated that A. orobius might be a junior synomym to A. costatus, which would bring the genus back to the monotypic status.

Material and methods
All studied specimens are preserved in 75% ethanol. Measurements and photographs were obtained with a Nikon DS-F2.5 camera mounted on a Nikon SMZ25 stereo microscope, using NIS-Elements Microscope Imaging Software with an Extended Depth of Focus (EDF) patch. Images were processed with Adobe Photoshop CS6 and assembled in Adobe InDesign CS6.
Terminology follows Hoffman (2008) and Enghoff (2017). Genus Attemsostreptus Verhoeff, 1941 Diagnosis Readily distinguishable from other genera of Spirostreptidae by the extremely long, slender lateral metaplical processes, in combination with the 'trachystreptinine' habitus, i.e., with marked longitudinal striae/keels over the entire body ring circumference. Krabbe (1982) presented a more detailed diagnosis, based mainly on male secondary sexual characters, as follows (translated from German and slightly modifi ed): collum without longitudinal keels. First pair of legs with strong 'shoulders'; the outer margin of the coxosternum smooth with 4 lateral small setae; prefemoral processes divergent, short, triangular with gently rounded margins, prefemora and prefemoral processes with pores. Gonopod sternum low and small; coxae slender, proplica (pp) simple, straight, bearing a few setae distally; metaplicae (mp) apically round to angular with extremely elongated lateral processes (lap) tapering into a claw-like apex, distally round to subrectangular; telopodite slender with a long, thin antetorsal process (atp), distal to (atp) becoming even slenderer, apex bifurcate with a thin serrated lamella and an acuminate apical process.

Attemsostreptus costatus
TELOPODITE. With a thin and curved antetorsal process (atp) ending in a pointed apex, post-torsal part forming a loop at mid-length, narrowing towards apex, apex bifurcate with a thin serrated lamella (sl) and an acuminate apical process.

Diagnosis
A species of genus Attemsostreptus that differs from A. costatus in the shape of the coxal apical and lateral processes, and in the presence of coxal distomesal spines.

Etymology
The species epithet is a Latin adjective meaning 'turned back' and referring to the shape of the lateroapical metaplical process of the gonopod.
Other material examined TANZANIA • 1 ♂, without data, mesal coxal spines (mcs) short, originating more distally than in the type specimens, mesoapical metaplical process (map) more truncate than round, distolateral spine maybe not quite so extraordinarily long, only 51 PR vs 56 PR in females from type series, body length (L) 65 mm, vertical diameter (H) 4.0 mm; NHMD621669 • 1 ♂, HNHM diplo-1697, without data, exactly like type specimens except that (mcs) are broken, 53 PR, L 70 mm, H 4.3 mm. COLOUR (Figs 2-3). After several years in ethanol, orange-reddish-brown, colour more intense at margin of the posterior half of metazonites, prozonites lighter with a hint of chestnut at the lateral margins. Head dark brown in occipital area, sputtered with yellowish on frons and reddish to yellowish brown on labrum and gnathochilarium. Antenna and legs yellowish. HEAD (Fig. 2). Four supralabral setae. Antennae reaching back to body ring 6. Eyes reaching slightly beyond mesal margin of antennal socket, each with ca 10 vertical rows of ommatidia, ca 5 horizontal rows, ca 35 ommatidia. Mandible as in Tropostreptus (Enghoff 2017): stipes in males with a small apicoventral lobe; odontomere long, moveable; sectile edge of psectromere with ca 5 lobes; ca 12 pectinate lamellae; one wide molar furrow. Gnathochilarium: prementum simple, not depressed; mentum with deep depression basally delimited by sharp ridge, as in Pseudotibiozus (Enghoff & Larsson 2018: fi g. 3); stipites with basal row of setae adjacent to mentum, apicolaterally with fi eld of many setae, distally with swelling harbouring one modifi ed seta, as in Pseudotibiozus (Enghoff & Larsson 2018: fi g. 3). COLLUM (Fig. 2). With protruding antero-ventral lobes in males. Three or four lateral furrows. No antennal groove.

MEASUREMENTS. Holotype
BODY RINGS (Figs 2B,3A,E). Prozonae with non-consipicuous furrows. Metazonae with regular longitudinal striae on entire body ring circumference, areas between striae elevated as ridges, ridges projecting beyond posterior margin of ring (Fig. 3B). Ozopore round, visible as bright spot rimmed with grey at the middle of metazona, behind suture (Fig. 3A), latter thin and sometimes curving at level of ozopore. Pleurotergal lobes meeting behind posterior pair of legs on each body ring.
COXA. Proplica (pp) simple, straight and broad, distomesal corner sharp, distal margin almost straight, oblique, surface in distal part with scattered bristles. Metaplica (mp) broad and stout; mesal margin straight, in basal part folded laterad as broad duplicature, subapically with large distad spine (mcs); in some specimens, (mcs) smaller and inserted more distally than in the holotype; metaplica apically gently rounded, mesoapical metaplical process (map) protruding as narrow ovoid mesal lobe, truncate in a few specimens (Fig 5B), meeting medially and totally lacking the meso-apical incision observed in A. costatus, laterally showing a strong constriction and protruding as lateroapical metaplical process (lap). This process (lap) very long, and strongly curved, fi rst directed laterad, then making a U-turn and pointing distomesad to mesad in most specimens, ending in claw-like pointed apex.
TELOPODITE. With a thin and curved antetorsal process (atp) ending in a pointed apex, post-torsal part long, with several torsions and an incomplete loop, thereafter narrowing towards apex and bifurcating into a thin serrated lamella (sl) and an acuminate apical process.

Discussion
Attemsostreptus was placed in "Trachystreptini" by Verhoeff (1941) who, however, questioned this assignment because although the body ring sculpture of Attemsostreptus suggested Trachystreptini, the gonopods of A. costatus in Verhoeff's view are more similar to those of Spirostreptus (in the then prevalent sense, cf. Hoffman et al. 2001) than to those of the then best known trachystreptinine genus, Lophostreptus Cook, 1895. Demange & Mauriès (1975) analysed the trachystreptinine spirostreptids in great detail and provided a revised classifi cation. They discussed the problems with delimiting Trachystreptini because the 'diagnostic' metazonital keels are developed to very different degrees. Demange & Mauriès (1975) mentioned another character which they regarded as a valid character for Trachystreptini, viz the shape of the lateral lobes of the collum. According to these authors, a part of the anterior surface of the collar lobes is fl exed towards the head capsule, thereby forming a cavity for accommodating the antennae. Hoffman (1980: 85) expressed general skepticism towards the internal divison of Spirostreptidae into subfamilies/tribes, and Krabbe (1982) like Verhoeff (1941) placed the name Trachystreptini in quotes, referred to it is a "Formengruppe" (morpho-group) and implicitly argued against its monophyly. Based on personal scrutiny of original literature as well as specimens from NHMW and NHMD, we can confi rm that the collum character emphasized by Demange & Mauriès (1975) is not present in all "trachystreptinine" genera. Whereas in some species, e.g., Furcillogonus frigidus Demange &Mauriès, 1975, andCalostreptus chelys Cook, 1895, the antennal grooves are very prominent, grooves are completely absent in the genus Attemsotreptus, as they are in Lophostreptus ptilostreptoides Carl, 1909, L. armatus Pocock, 1896, Anastreptus scalatus (Karsch, 1881 and Bucinogonus kandti (Carl, 1909). An antennal groove is present in Humilistreptus Demange, 1958(= Guviogonus Demange & Mauriès, 1975, a genus which includes a species with a "trachystreptinine" habitus as well as some without (Krabbe 1982).
If the relationships of Attemsostreptus to other spirostreptids thus remain unclear, the genus as such is well characterised and may be regarded as monophyletic based on the remarkable, very long and clearly apomorphic lateroapical metaplical processes.
The only known congener of A. refl exus, viz A. costatus, is also known only from the Uluguru Mts, suggesting a very restricted distribution area of the genus Attemsostreptus.