Rhinocoeta namaqua sp. nov. is recognised as a separate species from its closest relative, R. cornuta (Fabricius, 1781) after a review and close analysis of specimens recently collected in the semiarid region of the Northern Cape Province, South Africa. The new species can be readily separated from R. cornuta by the drastically reduced tubercle and associated depression on its anterior pronotal margin, particularly in the male. In addition, the general body shape of R. namaqua is more globose than that of R. cornuta, its average total length is larger, and its elytral costae are generally reduced and poorly visible, particularly at the level of the umbones. These characteristics make it practically impossible to separate the two sexes of R. namaqua, without inspection of the internal reproductive organs, as their external morphologies are virtually identical, unlike in R. cornuta. Finally, the aedeagal parameres of R. namaqua exhibit a narrower apex than those of R. cornuta and, in particular, lack the subapical hook-shaped lateral expansions that are so typical of all the other Rhinocoeta s. str. species. The new species appears to be restricted to specific bioregions of the Succulent and Nama Karoo biomes of the Northern Cape, and like all other species of the genus is generally found on or under mammal herbivore dung. Adult activity is limited to short periods immediately after rainfall events, during which individuals fly around and mate, but do not feed on either fruits or flowers.

Afrotropical region, Cetoniinae, new species, Northern Cape, Rhinocoeta, Xiphoscelidina

The genus Rhinocoeta Burmeister, 1842 currently includes two subgenera, the nominal one with five recognised species and R. (Haematonotus) Kraatz, 1880 with three species (Holm and Marais 1992, Beinhundner 2017). On the basis of adult and larval morphology, it has been argued that the genus may phylogenetically be placed close to the subtribe Xiphoscelidina rather than the Cetoniina, and that it constitutes part of a relict lineage derived directly from the most primitive Cetoniinae (Krikken 1984, Holm 1992, Smith et al. 1998). However, preliminary genetic DNA analyses seem to indicate that it may actually be closer to the more modern Cetoniina than previously believed (Kouklík 2017, Šípek et al. 2016). Consequently, their perceived “relictual/primitive” characters may actually represent adaptations to their peculiar habitat (e.g., droughts, extreme seasonality, scarce vegetation) and to their unique lifestyle, characterized by short bursts of activity and inability to feed at the adult stage (Smith et al. 1998).

Close analysis of a series of specimens collected during the past two decades in the Namaqualand and Upper Karoo regions of the Northern Cape, South Africa, has revealed that a new species previously confused with R. cornuta (Fabricius, 1781) occurs in these semiarid regions. None of the four synonyms used in the past to refer to R. cornuta can in fact be associated with the populations of the western part of the Northern Cape that constitute the new species. High resolution photos, collection data and circumstantial evidence obtained from the ZMUK (Kiel, Germany), BMNH (London, UK), MNHN (Paris, France), have led to the conclusion that Cetonia cornuta Fabricius, 1781, Scarabaeus arcas Olivier, 1789, Cetonia cornigera Gmelin (in L.), 1790 and Scarabaeus hispidolugubris Voet, 1779 all fit the typical characteristics of the Cape south-western populations of R. cornuta. This is also consistent with the period of their description, the late 18^th century, which coincided with the early exploration of the subcontinent. It would have been virtually impossible for the collectors of that period to have ventured beyond the colonial settlement of Cape Town and its immediate surroundings on the south and west coasts. Indeed, the remote Northern Cape populations remain very poorly sampled even in the modern era, with only approximately 20 specimens currently known for the new species here described.

Extensive observations made recently throughout the southern African region have also allowed for a better resolution of the distribution range and, particularly, the ecology of all the species of the nominal subgenus Rhinocoeta. This has prompted a reanalysis of the taxonomic position of R. limbaticollis (Péringuey, 1907), a rather enigmatic “species” described at the turn of the 19^th century on the basis of one female specimen only.

Specimens for this study were obtained through direct collections in the field during the period 1996–2018 (R Perissinotto and L Clennell legit), or from museum and private collections (as per list provided below). Fresh specimens were either caught in flight using standard nets after rainfall events, or collected on or under dung accumulations of a variety of herbivorous mammal species. Holotype, lectotype, paratypes, and other specimens of R. cornuta were analysed from high-resolution photographic material submitted by the museum curators listed in the Acknowledgement section.

For the description of morphological characters, the terminology used by Krikken (1984) and Holm and Marais (1992) is followed in this study. Specimen total length and maximum width were measured using a Vernier calliper, from the anterior margin of the clypeus to the apex of the pygidium and at the widest point of the elytra, respectively. Photos of specimen dorsal and ventral habitus were taken with a Nikon CoolPix S9700 digital camera with macro setting, while photos of the male genitalia were obtained using a Nikon DigitalSight DS-Fi2 camera attached to a Nikon SMZ25 dissecting microscope. The background was removed from the photos using Microsoft Word 2010 (Picture Tools), in order to increase clarity of resolution. The Combine ZP Image Stacking Software by Alan Hadley (alan@micropics.org.uk) was used to obtain z-stacking composite images.

The new description reported here brings to six the total number of species currently recognised within the nominal subgenus: R. sanguinipes (Gory & Percheron, 1833); R maraisi Holm, 1992; R. armata Boheman, 1860; R. limbaticollis (Péringuey, 1907), R cornuta (Fabricius, 1781) and R. namaqua sp. nov. (Holm 1992, Beinhundner 2017). There are then three further species in the subgenus Haematonotus Kraatz, 1880: R. (H.) turbida (Boheman, 1860), R. (H.) hauseri (Kraatz, 1896) and R. (H.) leonardi Beinhundner, 2013 (Beinhundner 2017). These have adequately been dealt with and thoroughly illustrated recently by Beinhundner (2013, 2017).

The issue of R. limbaticollis, however, remains a complex and unresolved one. It was described under a different genus, Lipoclita, by Péringuey (1907) mainly on the basis of its uncharacteristic mouthparts. These were regarded as fundamental by Péringuey (1907), but have subsequently been downgraded by Holm (1992), on the basis that mouthparts are variable and regressive, particularly in taxa such as all the members of the genus Rhinocoeta s. l. that do not feed at the adult stage. A close re-analysis of its key features reveals that R. limbaticollis largely resembles a typical dark/black female of R. (H.) turbida particularly in terms of protibiae, clypeus, antennal clubs and mesosternal process (Figs 6A–D), yet differs from this in terms of two key characters. These are: 1) the presence of cretaceous spots on the pygidium (Fig. 6D; although these are very regressive and not as developed as in males R. (H.) turbida); and 2) the lack of dense and coarse (even rugose on the pronotum) sculpture on the dorsal aspect (Fig. 6A).

Figure 6. 

Rhinocoeta limbaticollis (Péringuey, 1907), Holotype ♀ A dorsal habitus B lateral view C head and pronotum D pygidium. Photographs by Aisha Mayekiso; copyright Iziko Museums of South Africa.

Considering that this is unfortunately still the only specimen known for this “species”, despite the extensive searches that were undertaken in the area of the type locality during the past 30 years, there seem to be only two options left regarding its identity. The first is that it indeed represents a separate species to R. (H.) turbida, very rare or even extinct, as suggested by Holm and Marais (1992). This is difficult to believe, because the broader area of its type locality has not been completely “transformed” by mining and agriculture. There are in fact still protected and virtually pristine reserves in that area (e.g., Vaalbos/Mokala National Park, Sandveld Nature Reserve, Vredefort Dome World Heritage Site). Also direct observations made on virtually all the species of this genus indicate that they are very adaptable and resilient to anthropogenic activities (pers. obs.). The second option is that the specimen represents a case of gynandromorphy of R. (H.) turbida, with a mixture of female (e.g., bidentate protibiae, convex abdomen) and male characters (small and scattered sculpture, cretaceous spots on pygidium). It is hoped that the formulation of these working hypotheses may trigger further research, both in the field and the lab, to finally resolve this issues with conclusive evidence. To facilitate this, high-resolution photos of the specimen are presented here for the first time, courtesy of ISAM Collections Manager, Aisha Mayekiso.

Since the revision of the genus Rhinocoeta by Holm (1992) and Holm and Marais (1992), much has been learned about the biology/ecology and geographic distribution of the five confirmed species now constituting the nominal subgenus. Smith et al. (1998) reported that adults of the genus are generally collected under dung pads of cattle or in dung middens of a large variety of indigenous antelopes. In the Winterberg range of the Eastern Cape, larvae, pupae, and eclosing adults of R. sanguinipes have been found under a bontebok (Damaliscus dorcas dorcas) dung midden (Smith et al. 1998). Still in the Eastern Cape, but in the Sneeuberg, adults of R. maraisi sdults were collected in cattle dung (Holm and Stobbia 1995).

More comprehensive investigations carried out in the last 20 years have revealed that the four larger species, i.e., R. sanguinipes, R. maraisi, R. cornuta and R. namaqua, all depend on herbivorous and insectivorous mammal dung for their development (pers. obs.). On the other hand, the larva of R. armata has been observed making subsurface tunnels in sandy soil, in order to drag detrital matter underground, including leaf litter and pieces of dung pellets, of kudu (Tragelaphus strepsiceros) for instance (pers. obs., P Malec and P Šípek, pers. comm.). Rhinocoeta sanguinipes has been observed most frequently in large dung accumulations of klipspringer (Oreotragus oreotragus), but also in smaller dung deposits of aardvark (Orycteropus afer), red hartebeest (Alcelaphus buselaphus caama) and even domesticated goats, sheep and cows. Rhinocoeta maraisi seems to prefer dung droppings of mountain reedbuck (Redunca fulvorufula), Cape hare (Lepus capensis) and farmed horses and cows. Rhinocoeta cornuta on the Western and Eastern Cape south coast thrives on dung of farmed goats and sheep, but also of bontebok, hartebeest and other unidentified antelopes. Finally, the new species, R. namaqua, has so far been observed only on dung droppings of kudu and farmed sheep (pers. obs., J Ball and AP Marais, pers. comm.).

In terms of distribution, further to the ranges already reported in Holm (1992), Holm and Marais (1992) and Holm and Stobbia (1995), new data show that, apart from the Sneeuberge, R. maraisi occurs also in other major mountains of the Eastern and Northern Cape Karoo, such as the Bamboesberg, the Kikvorsberg and the Groot Tafelberg. The more widespread distribution of R. sanguinipes is confirmed, with populations occurring across the south-western and central parts of South Africa, as well as in southern Namibia. As previously reported by Holm (1992), the Namibian population exhibits entirely black ventral and dorsal habitus, and it has now been established that the westernmost South African populations also follow this colour pattern, generally lacking the typical brick-red pigmentation on ventral surface, legs, pygidium and pronotal lateral margins. Concerning R. armata, however, no evidence has been obtained in confirmation of the dubious distribution records of this species in Zimbabwe and Namibia, thus suggesting that the old references to these localities may be incorrect, as proposed earlier by Holm and Marais (1992). Finally, R. cornuta and R. namaqua seem to be clearly separated in their distribution by the Great South African Escarpment, with the first species restricted to the coastal lowlands and Cape Fold Mountains below it, while the second species has so far only been recorded in the north-western highlands above the escarpment (Fig. 5).

I would like to thank the following museum curators, researchers and owners of private collections for kindly providing photos, data and material for analysis: Aisha Mayekiso (ISAM), Ruth Müller (TMSA), Michael Kuhlmann (ZMUK), Max Barclay (BMNH), Alain Drumont & Florence Trus (ISNB), Riaan Stals & Werner Strumpher (SANC), Ashley Kirk-Spriggs & Burgert Muller (BMSA), Olivier Montreuil & Antoine Mantilleri (MNHN), Jonathan Ball & Andre Marais (BMPC), Gerhard Beinhundner (GBPC) and Thierry Garnier (TGPC). I am also very grateful to Lynette Clennell for organising all the photos and illustrations included in this work. The Nelson Mandela University (Port Elizabeth, South Africa) is thanked for providing facilities and funding towards the completion of his work. The Northern Cape Department of Environment and Nature Conservation is thanked for providing permits to collect specimens throughout the Province.