Ropalidia capensis ( DE SAUSSURE 1862 )

Ropalidia capensis ( DE SAUSSURE 1862) View in CoL

Icaria capensis DE SAUSSURE 1862

Type material. A female specimen ( RMNH. INS 108877 About RMNH ; NNM), with a label “ Pr.B.Sp. ”, denoting Cape of Good Hope, is currently labelled as the holotype specimen. One label on the type states “ TYPUS Ropalidia capensis ♀ Sauss ”, which is confusing as de Saussure used the genus name Icaria , suggesting that this label might have been added later on. This was already concluded by Kojima and van Achterberg (J. Kojima & Van Achterberg, 1997), who reported not finding any more type specimens and assigned lectotype status to this specimen .

The lectotype is a very lightly coloured specimen of this taxon, which shares several features with another very similar species, R. antennata (DE SAUSSURE) . The split between these two taxa was already discussed before by Kojima, who considered them as separate species, but misidentified them (J. Kojima, 2001). The final decision in this study was based on analysis of morphology, colour pattern and genetic data, enabling the separation of two closely related and similar taxa.

Comments. This is the largest member of the capensis -group, which seems to be geographically confined to South Africa and Namibia. The main features of females include elongated and straighter upes, horizontal or downward directing setae in the mid-section of the propodeal excavation and parallel-sided mandible; in addition to an overall darker and more diverse colour pattern, propodeum is always black. Taxonomic problems are very likely in separation from R. antennata (DE SAUSSURE) , especially in central and eastern parts of South Africa, where these two species are allopatric. Males are easy to separate based on the antennal morphology, the origin and shape of tyloids, and the terminal flagellomere shape.

Distribution. South Africa, Namibia. The distribution pattern of the entire capensis -group is interesting, primarily reflecting a lack of substantial overlap. The only exception to this is R. antennata (DE SAUSSURE) , which seems to span a large region and therefore is in contact with many species. The remaining species seem to have much narrower distributions, in some cases very narrow, suggesting likely ecological competition between them.

Genetics. The use of the COI gene had shown an interesting and somewhat complex situation. Four separate lineages were identified, three neighbouring and a more distant fourth one. This result suggests a somewhat complex situation that did not closely follow the geographical distribution. Furthermore, there were only minor colour differences between specimens from these clusters, further adding to the complexity of this situation. Despite these differences, all clusters were retained as a single species due to the lack of reliable morphological features.

When the entire capensis -group genetic data are analysed, three large clusters emerge, with two more lineages that consist of one or two species (Supplementary Figure 23 View FIGURE 23 ). The first large cluster includes five species ( R. capensis DE SAUSSURE , R. antennata DE SAUSSURE , R. luculenta sp. nov., R. nubila sp. nov. and R. tajiri sp. nov.). Shared morphological features are primarily seen in females, with the upward direction of setae of propodeal excavation, thinner mandibles, and slightly curved upes. Males do not seem to share morphological features, especially in comparison of R. luculenta sp. nov. with the remaining members of this cluster; males of this species by far resemble R. valentula sp. nov. Therefore, the genetic cluster membership seems to be primarily driven by female morphology, a previously reported finding for Palearctic Polistes (Schmid-Egger, van Achterberg, Neumeyer, Moriniere, & Schmidt, 2017) .

The second cluster corresponds to three newly described species, R. valentula sp. nov., R. nigrocerasina sp. nov. and R. retromaculata sp. nov. This cluster is characterized by substantial similarity in females (with limited morphological differences except the colour pattern), and minimal differences in males. The third cluster comprises four species, with R. kitui sp. nov. having a more basal and separate membership, while three more species seem to be more closely related to one another; R. novissima GIORDANI SOIKA , R. macloutsie sp. nov. and R. mosichi sp. nov. While R. kitui sp. nov. seems to be a morphological outlier indeed (with female clypeus and propodeum morphology providing sufficient elements for reliable determination), the remaining three species share numerous morphological similarities that confirm their relatedness (weakly develop juxtamandibular lobe in females and elongated antenna in both sexes). Lastly, two clusters remain; R. acuminata sp. nov. and R. ophuzi sp. nov. share membership, which corresponds well with their morphology (deep juxtamandibular excavations in females; males are unknown). Finally, isolated membership remains for R. makore sp. nov., which morphologically seems hard to compare, as it has features that cannot be easily attributed to any of these clusters, nor can it be aligned to any other species. Primarily, this is due to the strongly developed apical spot of the fore wing, weakly developed juxtamandibular lobes, elongated T1 and short AF2.