Description of Culicoides truuskae sp. n. (Diptera: Ceratopogonidae) from southern Africa

Culicoides truuskae Labuschagne and Meiswinkel sp. n. is described and illustrated in both sexes from material collected in South Africa and Namibia. It is restricted to the xeric western margin of the subcontinent, occurring in Fynbos, Nama-Karoo and Succulent Karoo ecoregions in South Africa and Desert and Savanna ecoregions in Namibia experiencing < 600 mm of rainfall annually. Culicoides truuskae sp. n. is part of the Afrotropical ‘plain-wing’ Culicoides in which the wing lacks a distinguishing pattern of light and dark spots; the diagnostic dark smudge that traverses wing cell r3 may result in C. truuskae sp. n. being misidentified as the sympatric but phyletically unrelated Culicoides herero (Enderlein) – (of the Similis group, subgenus Oecacta Poey). Additionally, this study is the first description of the male of C. herero. C. truuskae sp. n. and Culicoides coarctatus Clastrier and Wirth share similar characters in the male genitalia, although the two species are separable on wing pattern and female flagellum sensilla coeloconica (SCo) distribution. The breeding habitat and adult female blood-feeding preferences of C. truuskae sp. n. are not known. A maximum likelihood phylogenetic tree, using mitochondrial cytochrome c oxidase I (COI) sequence data, is provided to further clarify the relationship between C. truuskae sp. n., C. coarctatus and C. herero. Extensive light trap data, collected over 30 years, are used to map the distribution ranges of C. truuskae sp. n., C. coarctatus and C. herero in Southern Africa. Contribution The description of this new species and the description of the male of C. herero increases our understanding of the diversity and distribution of Culicoides species in southern Africa.


Introduction
In the genus Culicoides Latreille, the adult females of nearly all of the world's approximately 1350 species are blood feeders; as a result, many species transmit viruses, protozoans and filarial nematodes to an array of vertebrates including humans and their domesticated livestock and wildlife (Borkent & Dominiak 2020;Meiswinkel, Venter & Nevill 2004;Mellor, Boorman & Baylis 2000). A number of the viruses transmitted by Culicoides are of veterinary significance, particularly in Africa, where the most important are the causative agents of African horse sickness, equine encephalosis and bluetongue of sheep.
The Culicoides wing bears a '… characteristic pattern of dark and light spots … and is of primary importance in classification; however, in some groups or species it is poorly developed or even absent' (Blanton & Wirth 1979). As a rule, the wings of Culicoides from the forested tropics are far more brightly patterned (picture-wing) than the more dully coloured ones (plain-wing and halfwing) that predominate in drier and more open habitats. Of the 160 species that comprise the Afrotropical checklist, 17 are 'plain-wing' species, while an almost equal number possess a pattern that is 'poorly developed' and restricted to the basal half of the wing; these are referred to collectively as the 'half-wing' group, while the remainder possess a distinct pattern (Meiswinkel et al. 2004). As a result, more than 20% of Africa's biting midge species are difficult to identify accurately based on wing pattern alone; accordingly, these species are slide-mounted in order to examine other distinguishing features at higher magnification (Meiswinkel 1995).
Wing pattern reduction or its total loss is more pronounced in the xeric ecoregions of the world; for example, more than one-third of the 34 species of Culicoides recorded from the Arabian Culicoides truuskae Labuschagne and Meiswinkel sp. n. is described and illustrated in both sexes from material collected in South Africa and Namibia. It is restricted to the xeric western margin of the subcontinent, occurring in Fynbos, Nama-Karoo and Succulent Karoo ecoregions in South Africa and Desert and Savanna ecoregions in Namibia experiencing < 600 mm of rainfall annually. Culicoides truuskae sp. n. is part of the Afrotropical 'plain-wing' Culicoides in which the wing lacks a distinguishing pattern of light and dark spots; the diagnostic dark smudge that traverses wing cell r3 may result in C. truuskae sp. n. being misidentified as the sympatric but phyletically unrelated Culicoides herero (Enderlein) -(of the Similis group, subgenus Oecacta Poey). Additionally, this study is the first description of the male of C. herero. C. truuskae sp. n. and Culicoides coarctatus Clastrier and Wirth share similar characters in the male genitalia, although the two species are separable on wing pattern and female flagellum sensilla coeloconica (SCo) distribution. The breeding habitat and adult female blood-feeding preferences of C. truuskae sp. n. are not known. A maximum likelihood phylogenetic tree, using mitochondrial cytochrome c oxidase I (COI) sequence data, is provided to further clarify the relationship between C. truuskae sp. n., C. coarctatus and C. herero. Extensive light trap data, collected over 30 years, are used to map the distribution ranges of C. truuskae sp. n., C. coarctatus and C. herero in Southern Africa.
http://www.ojvr.org Open Access Peninsula (Boorman 1989) comprise 'plain-wing' or 'halfwing' species. Globally, deserts form the core of many of the world's drier tracts, with characteristic unstable, highly labile entities that expand and contract as climatic conditions vary. In Africa's past during drier epochs, 'arid corridors' developed and expanded, cutting diagonally across the waist of the continent. This facilitated faunal and floral interchange between the deserts and xeric shrublands of the two 'corners' of the continent, namely the Horn of Africa and the Karoo-Namib. In mammals, the best examples of the existence of this ancient corridor are the oryx, dik-dik and bat-eared fox, while in birds it is among the bustards, larks, and smaller hornbills (Kingdon 1990). This interchange also likely involved insects. These past events complicate taxonomic studies because sister taxa, separated by vast distances and across timeframes of unknown length, may be common.
As observed, the taxonomy of the Culicoides of the Arabian Peninsula has enjoyed some attention (Boorman 1989), in contrast to the Culcoides species 5000 km away in the southwestern corner of Africa. This region is characterised by deserts and xeric shrublands that have long prevailed on the broad strip of land that runs for almost 2000 km along the Atlantic seaboard, an area of great dryness the early Khoisan peoples called 'Karoo'. Phytogeographers have parceled this 'great dryness' into a number of terrestrial ecoregions; the one from which nearly all collections of Culicoides truuskae sp. n. were made is the 'Succulent Karoo' (ecoregion 110; Burgess et al. 2006) and recognised to be the '… world's only plant hotspot that is entirely arid' (Burgess et al. 2006). Explosive speciation among succulents and geophytes within the larger Karoo-Namib phytochorion (Van Wyk & Smith 2001) likely accounts for the co-existence of a rich complement of arthropod endemics, especially among the arachnids, hopliniid beetles, aculeate Hymenoptera, and reptiles (Burgess et al. 2006). Hopliniid monkey beetles, masarine wasps and colletid, fideliid, and melittid bees serve as important pollinators (Burgess et al. 2006). In sharp contrast, the bird (one strict endemic) and mammal faunas (three endemics, of which two are subterranean) are depauperate (Burgess et al. 2006).
Reduced vertebrate diversity will impinge greatly upon the feeding opportunities available to blood sucking insects such as Culicoides. However, it is equally valid to argue that sustained aridity, across evolutionary time, is strongly selective and so bound to have given rise to a specialised, albeit greatly reduced, biting midge fauna within the arid heart of southwestern Africa. This possibility remains unexplored.
In this study, the authors describe C. truuskae sp. n. from the western deserts and xeric shrublands of South Africa and Namibia. In addition, the male of Culicoides herero is described for the first time. At the species level, C. truuskae sp. n. is compared taxonomically with Culicoides coarctatus (because of shared characters of the male genitalia) and to C. herero (with which the new species may be confused owing to wing pattern similarity). The authors assess the higher taxonomic position of the three. Finally, extensive light trap data collected over 30 years allowed the authors to map in detail the distribution of the three species within southern Africa and to provide some data on the seasonal phenology of its constituent species.

Material and methods
All specimens of C. truuskae sp. n. examined were collected with the Onderstepoort Veterinary Institute (OVI) black light suction trap. A small portion of this material is slidemounted in Canada balsam and housed in the Insect collection of the Agricultural Research Council -Onderstepoort Veterinary Institute (ARC-OVI). Specimens were examined using a Zeiss Axiostar compound microscope equipped with an ocular reticule for measurements. All measurements are in micrometres (μm), except wing lengths (mm). Photographs were taken using a Zeiss AxioCam ERc5s camera, measurements were carried out in ZEN 2.3 lite and the images then enhanced in Adobe Photoshop (elements 3.0). In general, the descriptive taxonomic terminology follows Khamala and Kettle (1971) and Boorman (1989), except for (1) the flagellar segments numbered 1-13 (instead of referring to antennal segments 3-15) and (2) the proboscis/head (P/H) ratio is used instead of the inverse (H/P). P is the distance from the tip of the proboscis to the tormae and H the distance from the tormae to the insertion alveolus of the interocular seta. Terms for the various types of sensilla situated on the flagellum follow Meiswinkel (1995).
Total deoxyribonucleic acid (DNA) extracts were obtained from individual Culicoides using the ZR Tissue and Insect DNA Miniprep kit (Zymo Research) and following the manufacturer's instructions. The wings, head and antennae were dissected off and mounted onto a microscope slide as a reference sample for each specimen before destructive DNA extraction was performed for female specimens. For male specimens, the wings and the genitalia were dissected of and mounted before extraction. A 700 base pairs (bp) region of the mitochondrial cytochrome c oxidase I (COI) gene amplified using the primers LC01490: GGTCAACAAATCATAAAGATATTGG and HC02198: TAAACTTCAGGGTGACCAAAAAATCA as previously described (Folmer et al. 1994 ). An alignment containing all COI consensus sequences, as well as selected reference sequences obtained from GenBank was constructed in Mega X using Clustal W (Tamura et al. 2013). The best fitting model was determined using Mega X, and a General Time Reversible model with Gamma distribution (4) and Invariant sites (G + I) was used to compute the Maximum Likelihood phylogeny with 1000 bootstrap replicates. Percentage sequence identity was determined between C. truuskae sp. n. sequences as well as between C. truuskae sp. n., C. coarctatus and C. herero and remaining sequences in the alignment using CLC Genomics Workbench 9.0.1. (www.CLC genomics.com).
The distribution map of C. truuskae sp. n, C. coarctatus and C. herero in southern Africa was generated using QGIS 22.7 software (QGIS Development Team 2022). The seasonal abundance and monthly prevalence of C. truuskae sp. n., C. coarctatus and C. herero is depicted in a single Mondrian Matrix; the matrix consists of squares colour-coded according to the log10 (n + 1) transformed monthly average abundance of each species.

Ethical considerations
This article followed all ethical standards for research without direct contact with human or animal subjects.

Differential diagnosis
Medium sized, fresh specimens brown and slide mounted specimens light brown, female eyes narrowly separated. Sensilla coeloconica (SCo) was present on flagellomeres 1, 3, 5, 7, 9-12 in females and 1, 11-12 in males. The third palpal segment inflated, sensory pit wide and deep. Wing was without pale or dark markings and anterior half of cell r 3 was noticeably darkened. Spermathecae: two slightly unequal, ovoid spermathecae with a rudimentary third spermatheca present, sclerotised ring absent. Male genitalia: Ninth tergum with long, basally stout apirolateral processes; ninth sternum with broad and moderately deep caudomedian emargination, the ventral membrane spiculate; dorsal root of basistyle moderately short and stout, ventral root short; aedeagus triangular, narrowing gradually to broad apex, aedeagal arch moderately high; parameres broadly joined at the base mesally, each paramere slender, tapering to a single fine point.

Etymology
Named in honour of Dr Truuske Gerdes in recognition of her work on arboviral diseases of veterinary importance in South Africa.

Bionomics
The immature stages, larval habitat, blood-feeding habits and vector status of C. truuskae sp. n. remain undescribed.

Distribution
South Africa: Eastern, Northern and Western Cape provinces. Central and southern Namibia in: Erongo, Karas, Khomas and Otjozondjupa regions.

Differential diagnosis
Small sized, fresh specimens dark and slide mounted specimens light brown. Female eyes narrowly separated. Sensilla coeloconica present on flagellomeres 1-8 in females and 1, 6, 8-10 in males. The third palpal segment inflated, sensory pit wide and deep. Wing with indistinct 1st and 2nd costal spots present and a faint pale spot in cell cua1. Costal vein between the first and second costal spots dark. Spermathecae: two slightly unequal, ovoid spermathecae with rudimentary a third spermatheca present, and a sclerotised ring present. Male genitalia: ventral roots of basistyle with anterior processes connected by a very narrow hyaline membrane; shoulders of aedeagus with a pair of posteriorly directed, pointed processes; distal median process long and stout with truncate apex. Paramere with a large, anterolaterally directed basal knob; main stem straight, stout, with a large subapical ventral lobe; distal portion recurved to a slender, pointed apex, and with 5-6 lateral fringing spines.  Abdomen. Genitalia (Figure 4c, e, f). Gonocoxite signifi cantly broader basally, tapering distally; dorsal root long and stout, ventral roots long with anterior processes connected by a very narrow hyaline membrane, posterior heel better developed than the anterior toe ( Figure 3f). Paramere ( Figure 4e) with a large, anterolaterally directed basal knob; main stem straight, stout, with a large subapical ventral lobe, distal portion recurved to a slender, pointed apex, and with 5-6 lateral fringing spines. Aedeagus (Figure 4f) with a broad, deep basal arch, lateral arms slender, bases curving laterally; shoulders of aedeagus with a pair of posteriorly directed, pointed processes; distal median process long and stout, with truncate apex. Ninth sternum with shallow caudomedian excavation, ventral membrane not spiculate. Ninth tergum with short slender, pointed apicolateral processes (Figure 4c).

Bionomics
The immature stages, larval habitat, blood-feeding habits and vector status of C. herero remain undescribed.

Differential diagnosis
Medium-sized, fresh specimens dark brown and brown in slide mounted specimens. Female eyes moderately separated. Distal antennal segments not elongated, SCo on each of flagellomeres: 1, 3, 5, (6), 7, 9, 11, 12 in females and 1, 3, 5, 7, 11, 12 in males. Occasionally SCo absent on flagellomere 6 in females and on flagellomeres 3, 5 and 7 in males. The third segment of maxillary palpus moderately expanded, with a large, shallow sensory pit. Wings were patterned with a large pale spot over r-m crossvein extending anteriorly to costal margin and caudally into cell M2; cell r3 with a pale spot on anterior margin just distad of second radial cell; pale spots in distal portion of cells M1, M2, M4, and anal cell; cell M1 with a pale spot near base; cell M2 with a pale streak in basal half Thorax. Scutum brown in colour. Wing (see Figure 6a). Abdomen. Genitalia (Figure 6d-e) stout, square. Gonocoxite significantly broader basally, ventral root absent; dorsal root stout, moderately long and straight, tip broadly rounded. Gonostyle basal two-thirds broadly inflated, narrowing quite abruptly to sharply pointed tips. Parameres broadly fused in mid-line, the distal portion extending posteriad, each paramere tapering into long, slender, slightly bowed, inwardly curved, fine, erect point (Figure 6e). Aedeagus robust, triangular, wellpigmented, narrowing gradually to broad apex, aedeagal arch moderately high, slightly convex, basal feet well-developed, projecting laterad at right angles (Figure 6d). Sternum 9 with broad and moderately deep caudomedian excavation, posterior membrane rather densely spiculate, spiculation extending posteriad over basal arch of aedeagus ( Figure 6d). Tergum 9 trapezoidal, narrowing gradually, apicolateral processes moderately broadly separated each process short, broad basally; caudal margin of tergum with small, indistinct notch medianally.

Bionomics
The immature stages, larval habitat and vector status of C. coarctatus remain undescribed. Blood-feeding habits are http://www.ojvr.org Open Access unknown although females were collected feeding on humans in Chad (Glick 1990).

Sequencing and phylogenetic analysis
The phylogenetic relationship between C. truuskae sp. n., with other species were constructed based on the partial COI sequences, with Forcipomyia crassipes as an outgroup (Figure 7). Both males and females of C. truuskae sp. n., C. herero and C. coarctatus were sequenced. The phylogenetic analysis of all three species clearly indicate that the males and females belong to the same species, with more than 97% sequence similarity within the three groups. An average of 88% sequence similarity was observed between the sequences of C. truuskae sp. n. and either C. herero or C. coarctatus species, while all the other sequences shared less than 84% sequence similarity (Table 1). Sequences from 11 additional species were obtained from Genbank and included into the phylogenetic analysis. The related species clustered together in three distinct clades, that is subgenera Avaritia, Culicoides as well as the Milnei group.
The clustering of C. truuskae sp. n. and C. coarctatus together is weakly supported with a bootstrap value of 51, and both species cluster weakly with C. herero with a bootstrap value of 35. In the mean genetic distances, C. truuskae sp. n. share the highest percentage sequence identity with C. herero and C. coarctatus compared with the other species (Table 1). In order to establish the relatedness of these three species, additional nuclear gene regions will need to be sequenced.

Discussion and conclusion
In C. truuskae sp. n., the charcoal-coloured smudge broadly running the length of cell r3 is diagnostic and are similar to the one seen in the wing of C. herero. In his original description of C. herero, based on a female from Namibia, Enderlein (1908) singled out this smudge and, in German, referred to it as a 'verwaschene Fleck' (vF). Translated, this refers to something as 'washed out, faded, watery, blurred'; these nuanced meanings aptly capture the nature of the 'Fleck' and, although obvious, its margins are indistinct rather than sharply defined. The vF is found on the wing in both C. herero and C. truuskae sp. n. and because these two species occur in sympatry, raises the point as to which species Enderlein was dealing with when he briefly described C. herero over a century ago. Although long considered 'lost', the discovery of the holotype in the holdings of the Berlin Museum enabled Szadziewski (1986) to re-examine it; along with a full description, he also provided a figure of the wing clearly showing the vF as well  as the first and second costal spots. In contrast with the single female that Szadziewski (1986) examined, the majority of our specimens had a very indistinct pale spot in cell cua1. Unfortunately, the Namibian samples available were collected 250 km from Rooibank where the holotype was collected. As to its taxonomic position, Szadziewski concluded, correctly, that C. herero is a member of the Similis species group (one of the many clades that comprise the subgenus Oecacta Poey). Szadziewski's full redescription of the holotype of C. herero makes it clear that Enderlein did not have C. truuskae sp. n. before him all those years ago (Szadziewski 1986).
After a 2-year weekly light-trap survey conducted over 30 months (between 1996 and 1999) at 40 livestock holdings distributed across South Africa, C. coarctatus was found at 21, C. herero at 17 and C. truuskae sp. n at 9 of the sites. Monthly abundance data for C. coarctatus was only available for 18 sites, C. herero for 8 sites and C. truuskae sp. n for 7 sites.
To highlight similarities and differences between C truuskae sp. n., C. coarctatus and C. herero, the monthly prevalence data are presented in the Mondrian matrix ( Figure 8 The respective geographic ranges of C. truuskae sp. n. and C. coarctatus verge on allopatry (Figure 9). Since its description almost 60 years ago, the seasonal phenology of C. coarctatus has remained '... virtually unknown', in spite of its wide occurrence within the savannas of Africa, from Nigeria and Chad to Ethiopia in the north, down to Kenya, Zimbabwe and South Africa in the south (Glick 1990). All the sites for C. coarctatus were located in the eastern two-thirds of South Culicoides coarctatus and C. truuskae sp. n. have near allopatric ranges while C. herero and C. truuskae sp. n. have sympatric ranges. In sharp contrast to C. coarctatus, the geographical range of C. truuskae sp. n. is restricted to the northern, western and eastern Cape provinces, but this species is seldom found abundantly in light trap collections (Figure 9, Online Appendix 1 Table 2-A1). Culicoides truuskae sp. n. and C. herero occur in the Fynbos, Nama-Karoo and Succulent Karoo ecoregions in South Africa, although C. herero also occur in the southern savanna ecoregion of the eastern and western Cape provinces and both occur in the desert and savanna ecoregions in Namibia. Little is known about C. truuskae sp. n. abundance, prevalence and seasonal distribution in Namibia owing to limited sampling studies. Culicoides coarctatus distribution is wider in all ecoregions except the Succulent Karoo and Desert ecoregions into areas with up to 1500 mm rainfall annually. Culicoides herero has a wider distribution than C. truuskae sp. n. and penetrates into areas with up to 1000 mm rainfall annually (Figure 9, Online Appendix 1 Table 2-A1).
Culicoides truuskae sp. n. and C. coarctatus can be distinguished based on their size, wing pattern and SCo distribution on the flagellomeres (Table 2). Culicoides truuskae sp. n. is a medium sized species, lacking wing maculation pattern, with SCo on flagellomeres 1, 3, 5, 7, 9-12, while C. coarctatus is a much smaller species, with a definite wing maculation pattern (pale spots at apex of the wing margin, with two rows of spots from the first and the second costal spots to the posterior wing margin) and SCo on flagellomeres 1, 3, 5, (6), 9, 11-12. The male genitalia of C. truuskae sp. n. and C. coarctatus are similar, that is fused parameres and the densely spiculate posterior membrane of sternum 9, with exception of the aedeagal arch that is high and slightly rounded in C. truuskae sp. n., but only moderately high and rounded in C. coarctatus.
Culicoides herero can be distinguished from C. truuskae sp. n. and C. coarctatus based on size, wing pattern and SCo distribution on the flagellomeres (Table 2). Culicoides herero is a small species lacking wing maculation pattern except for very indistinct pale spot in cell cua1 and SCo on flagellomeres 1-8. The aedeagus and parameres of C. herero differ from both C. truuskae sp. n. and C. coarctatus. In C. herero the shoulders of aedeagus have a pair of posteriorly directed and the parameres has a large, anterolaterally directed basal knob; main stem straight, stout, with a large subapical ventral lobe, distal portion recurved to a slender, pointed apex and with 5-6 lateral fringing spines.
Culicoides confusus Carter, Ingram and Macfie, resemble C. coartcatus, in wing pattern, but differ in having SCo on flagellomeres 1, 3, 5-7, 9-12 and a mean lower palpal ratio of 1.93 versus 2.31 in C. coarctatus. The males of C. confusus have not officially been described, although Boorman and Dipeolu (1979) give a description of the male based on a single specimen. This male differs from C. coarctatus in the posterior membrane of sternum 9 being bare and that the parameres end in curved tips. In wing pattern C. clarkei Carter, Ingram and Macfie, also resemble C. coartcatus but differ in having SCo distribution on the flagellomeres 1-7, 9-12. The male genitalia also differ in having separated parameres and a triangular aedeagus with a long finger-like median process. These two species are currently not known to occur in South Africa.
In conclusion morphology, sequence results based on the phylogenetic analysis of the partial COI gene region support C. truuskae sp. n as a valid new species and the description of the male of C. herero for the first time.