Description of Two New Species of Zinophora Chamberlin, 1927 (Diplopoda: Spirostreptida: Harpagophoridae), with Discussion of Species Groups in the Genus

ABSTRACT The millipede genus Zinophora (Chamberlin, 1927) was previously revised based on the species present south of the Zambezi and Kunene Rivers. Since that revision, two additional new species have been discovered and are here described: Zinophora lobata (Mozambique) and Z. taromberai (Zimbabwe and Tanzania). This brings the total number of described species in the genus to 21. An updated key to species in the genus is presented. Detailed illustrations of the gonopods supplement the descriptions and a distribution map of Z. taromberai in Zimbabwe is provided. There appear to be up to three species groups in the genus and the character states defining these groups are discussed.


INTRODUCTION
The millipede family Harpagophoridae is not only found in Africa, but also occurs in Southeast Asia and the east Indies (Hoffman 1993), and is probably the most characteristic and conspicuous element of the millipede fauna of that region (Hoffman 1975). The family is represented in Africa by the following genera: Poratophilus Silvestri, 1897 (two species); Harpagophora Attems, 1909 (six species); Zinophora Chamberlin, 1927 (21 species); Apoctenophora Hoffman & Howell, 1982 (four species), and Metaphora Redman, 2003 (one species).
The most widely distributed and speciose genus of Harpagophoridae in Africa is Zino phora, which ranges from the Western Cape (South Africa) and Namibia in the southwest, to Mozambique, Zimbabwe, Zambia, Malawi and Tanzania in the east. It is largely restricted to the savannah biome, but some species are adapted to rainforest conditions (Hoffman 1993) and others to the grassland biome . The genus currently includes 19 named species , of which 16 are from Africa south of the Zambezi and Kunene rivers. Two species groups were identified by , the munda group and the gracilus group. It is anticipated that more millipede species will be discovered as more sampling is conducted (Mwabvu et al. 2007).
Issues relating to poor collecting effort are particularly relevant in the millipedes, and this is attested to by the discovery of the new species described in this paper. There is a need for a much stronger drive to document the biodiversity of southern Africa. Invertebrates perform essential ecosystem functions, yet they are largely overlooked in mainstream conservation planning (Huntly et al. 2005), largely as a consequence of generally poor taxonomic data. More data will certainly contribute to the inclusion of invertebrates in biodiversity conservation planning (Slotow & Hamer 2000). The challenge is to cover as much of the poorly sampled areas in Africa as possible to overcome collecting bias and, more importantly, to achieve this in the face of the current accelerated rate of development in Africa (Hamer & Slotow 2002).
The aim of this paper is to assess the taxonomic status of new material of Zinophora collected outside South Africa, and to describe two new species represented by this material. In addition, a revised key including known species of the genus is presented and the possibility of three species groups in the genus is discussed.

MATeRIAl AND MeTHODS
Material for this study was obtained from the Natal Museum (Pietermaritzburg, South Africa; NMSA) and the Virginia Museum of Natural History (Virginia, USA; VMNH).
Data for specimens examined are arranged in alphabetical order by country, followed by province, and then by date of collection, with the oldest specimen listed first. Information for each specimen lot (specimens from the same locality and with the same catalogue number) has been recorded as stated on the label.
The number of body rings (excluding the pre-anal ring) was counted, noting the position of the first ozopore. The anal valve was not counted as a segment. The following characters were examined: (1) the colour and texture of the head; (2) the colour and (C) Z. taromberai, telopodite, posterior view. Abbreviations: f -femoral spine, mt -metaplica, p -palette, pl -proplica, sl -spatulate lobe, t -thumb, ts -telocoxal spine. Scale bars = 1 mm. texture of the clypeus, noting the number of supralabral pits present; (3) the number of ocelli and horizontal rows of ocelli; (4) the anterior and posterior angle, submarginal groove, and colour of the collum; (5) the gnathochilarium and prebasillar plate of the gnathochilarium; (6) the mandibles; (7) the colour, structure and length of the antennae and legs (the average from segments 5, 24 and the last segment); (8) the structure of the first pair of male legs; (9) the structure of the gonopods; (10) the colour and texture of the prozonites, mesozonites and metazonites; (11) the colour, structure and texture of the pre-anal ring, anal valve, and caudal spine; (12) the margin of the limbus; and (13) the shape of the hypoproct. Colour was described from specimens preserved in 70 % ethanol since their collection. All descriptions and measurements are based on males, unless otherwise stated. Measurements are provided in the following sequence: type specimen; minimummaximum; [mean]. Means were only calculated when more than three specimens were measured. All measurements are in mm. Measurements of the legs, body and antenna were taken using Vernier callipers (range 0-200 mm). Body width measurements (horizontal diameter) were taken at the collum and at the widest trunk segment (between segments 20 and 31). For measurements of length, a specimen was straightened out and measured, and when this was not possible (as a result of specimens locked in a spiral position), a piece of soft wire was used to measure body length. In cases where the specimen was broken, the individual pieces were measured separately and these lengths were totalled.
The first pair of male legs and the gonopods were removed from the millipede and one of the telopodites was separated from the gonopod coxites. These structures were illustrated using a Wild Heerbrugg M5A stereo microscope and drawing tube, and photographed using auto montage software (leica Microscope MZ12s with 3 CCD Toshiba Camera).
The lack of appropriate and consistently used descriptive terminology providing a suitable basis for comparison has hindered the comprehensive use of the diverse forms of the genitalia in spirostreptid taxonomy (Hoffman 2008). Currently the male genitalia provide the vast majority of informative taxonomic characters in the Harpagophoridae and in other millipede groups (Pimvichai et al. 2009). More recently a considerable amount of effort has been made to address these inconsistencies, which at times suggest unestablished homologies and development patterns, with suggestions for refinement of some existing terminology (Hoffman 2008;Pimvichai et al. 2009). For purposes of continuity, the terminology used for millipede anatomy and gonopod characters follows a combination of Attems (1928), Hoffman (1994), Hamer (1999), and Van den Spiegel & Hoffman (2001). It also factors in more descriptive terminology recently proposed by Hoffman (2008) and Pimvichai et al. (2009) (Fig. 1).

TAXONOMy
Key to the species of Zinophora 1 Metaplica lobed and with edge of lobe extending beyond medial margin of proplica, or even overlapping opposite metaplica medially (as in Figs 2, 7). Thumb of telopodite represented by one or two processes; if one it is narrow and elongate or broad based and elongate, otherwise reduced. Telopodite with one femoral spine (as in Fig. 4) 4, 5 (Carl, 1917) (Attems, 1928) -Telocoxal spine narrow at base, smooth on apical surface and directed aborally, with apex directed towards proximal part of telocoxite. Two femoral spines angled at about 45°, with one spine strongly curved medially ...... thukela  Zinophora lobata Redman, sp. n. Figs 1A, 2-6 etymology: From latin lobata (lobed) adjective; refers to the elongated protuberance along the medial margin of the metaplica. Diagnosis: Metaplica triangular, with medial margin formed into extended finger-like lobe or medial process directed posteriorly (Fig. 2). Telocoxal spines curved with apexdirected basad (Fig. 3). Telopodite with one long and narrow acutely produced femoral spine (Fig. 4). Thumb narrow, curved in direction of palette and apically acute ( Anal valve: Surface bulbous and smooth. Margin smooth and raised, delimited by welldefined granular depression. First pair of male legs: Prefemora in contact medially. Syncoxosternum with a slit medially between paracoxites (Fig. 6). Gonopods: lateral margin of metaplica distally rounded and sloping diagonally over proplica, concealing telocoxal and femoral spines (Fig. 2). Short, downcurved medial telocoxal spine (Figs 2, 3). Metaplica triangular, with medial margin formed into extended finger-like lobe (Fig. 2). Telopodite with one long, narrow and acutely produced femoral spine concealed by metaplica (Figs 3, 4). Thumb narrow, curved in direction of palette and acutely produced. Spatulate lobe a thin hyaline plate broadening apically, with tip curled away from other apical elements (Fig. 5). Distribution: The only record for this species is from Pomene in the Inhambane province in southern Mozambique. Remarks: The thumb on the apical elements of the telopodite is narrow, curved in the direction of the palette and acutely produced, making this species a member of the Zinophora munda group. The locality of the species is also consistent with the distribution of other species in the Z. munda group, which is generally in the northern half of South Africa, extending beyond the limpopo River ).

Zinophora taromberai
Diagnosis: Apical surface of metaplica with prominent ridge extending to lateral margin. Medial margin of metaplica formed into posteriorly directed lobes (Figs 7, 8). Telocoxal spines medially directed with broad base. Femoral spine long, reaching the distal margin of the metaplica, and curving slightly with the lateral margin of the metaplica (Fig. 9). Thumb on apical elements of telopodite formed into two processes, one narrow and falcate and directed towards palette, the other wide and concave (Figs 9, 10). First pair of male legs: Prefemora in contact medially. Syncoxosternum with a slit me dially between paracoxites. Prefemoral process produced into long finger-like pro tuberances supported by the coxal shelf. Gonopods: Apical surface of metaplica with prominent ridge extending to lateral margin (Fig. 7A). Medial margin of metaplica formed into posteriorly-directed lobes (Fig. 7). Telocoxal spines medially directed, with broad base (Fig. 8). Telopodite with one long, narrow and acutely produced femoral spine partially concealed by metaplica with slight curvature (Fig. 9). Thumb on apical part of telopodite formed into two processes, one narrow and falcate and directed towards palette, the other wide and concave (Fig. 10). Spatulate lobe a thin hyaline plate broadening apically, with tip curled away from other apical elements (Fig. 10).
species, in that the apical surface of one of the telocoxal spines was dentate. A similar intraspecific deviation has been observed in other Zinophora species such as Z. junodi, Z. thukela and Z. punctata. However, in these species both the telocoxal spines were dentate, unlike in the Z. taromberai specimen, where only one of the telocoxal spines was dentate on the apical surface.

DISCUSSION
In all millipedes the male gonopods are known to offer useful taxonomic characters, and this is affirmed in Zinophora and in this study. In Zinophora the structure of the spatulate lobe, in combination with the general form of the distal margin of the telocoxites comprising a proplica and metaplica folded against each other, with the metaplica lobed mesiad and the proplica bearing a mesal projection or coxal spine, are the diagnostic characters of the genus. The structure of the spatulate lobe-one of the apical elements of the telopodite-is by far the most consistent character state distinguishing Zinophora from other genera of the Harpagophoridae and, together with several other character states, strongly defines the genus. The structure of the thumb, which is also located on the apical elements of the telopodite, is notable in distinguishing species groups.
In this study it became evident that some Zinophora species have a thumb on the apical region of the telopodite which is formed into a single finger-like projection (Zinophora munda group, 9 species), whereas some species have a single saucer-shaped laminate plate with a dentate margin (Zinophora gracilis group, 9 species), while in others it is represented by two processes which are a combination of the aforementioned character states. In Z. lobata the thumb is formed into a single finger-like projection, thereby making it a member of the Z. munda species group, however in species such as Z. knipperi, Z. distincta and Z. taromberai, a combination of the above mentioned character states (elongate finger-like process and saucer-shaped laminate plate) are observed. This suite of characters, together with the general shape of the telocoxite (i.e. comprising of a metaplica and proplica), suggests there may be three species groups in Zinophora.
The above distinctions have been based primarily on male genitalic morphological differences. We recommend that molecular investigations based on mitochondrial DNA be considered to provide further resolution by revealing the genetic diversity of the species groups. This could also clarify to what extent the apical elements of the telopodite in this genus provide accurate phylogenetic synapomorphies and what weight should be attached to the shape of spatulate lobe in distinguishing the genus, and the thumb in distinguishing the species groups. The morphological variation in the gonopod structure of the species groups could potentially be geographically associated but the phylogeographic basis for these species groups will be addressed in detail in a future paper.
Prior to this publication, only two species of Harpagophoridae were known from Mozambique (Z. brevilobata and Z. junodi), and a recent trip to the region resulted in the discovery of a new species (Z. lobata). A similar situation pertained in Zimbabwe, whence only one species was previously known (Z. munda) and a single field trip added another species with minimal collecting effort. This suggests that much more work in the region is needed to give a more accurate representation of regional millipede biodiversity, allowing more informed pronouncements on the conservation status of relevant taxa.