The genus Afrosyrphus Curran (Diptera, Syrphidae), with a description of a new species

The flower fly genus Afrosyrphus Curran, 1927 (Diptera, Syrphidae) is revised and a new species, Afrosyrphus schmuttereri sp. nov., from Kenya and Uganda is described. Diagnoses, illustrations, DNA barcodes and known distributional data are provided for the two species of this genus, as well as an identification key. A critical review of the published literature is also provided.

view. Wing length was measured from the wing tip to the basicosta. Photographs were composed using Zerene Stacker ® ver. 1.04 software (Richland, WA, USA), based on images of pinned specimens taken with a Canon EOS 7D ® mounted on a P-51 Cam-Lift (Dun Inc., VA, USA) and with the help of Adobe Lightroom ® ver. 5.6. Simple-Mappr (Shorthouse 2010) was used to create Fig. 2.

Adult identification
Existing available keys were used to identify the collected specimens. Curran (1927), Hull (1949) and Vockeroth (1969) were used to determine the genus; then, the specimens from Taita Hills were checked by direct comparison against material from several collections, including the holotype of A. varipes, and against the original description of A. varipes.

Molecular studies
One or two legs of several specimens were used for DNA extraction. Extractions were carried out using the DNeasy Blood and Tissue Kit (Qiagen Inc., Santa Clara, CA, USA) following the manufacturer's protocol. Entire specimens or remnants of specimens were preserved and labelled as DNA voucher specimens for the purpose of morphological studies and deposited in the above-mentioned collections, as listed in the 'Material examined' sections.
We ran a distance-based Neighbor-Joining analysis using the Jukes-Cantor Model as implemented in Geneious ver. 7.1.3, where we included several DNA barcode sequences for species of the genera Epistrophe and Epistrophella Dušek & Láska, 1967 in order to visualize the distance between the two species of Afrosyrphus. We also ran a maximum likelihood (ML) analysis using Garli ver. 2.01 (Zwickl 2006(Zwickl , 2011 and the GTR + I + G model (Fig. 6). Forty-eight independent runs were conducted using scorethreshforterm = 0.05 and significanttopochange = 0.0001 settings and the automated stopping criterion, terminating the search when the ln score remained constant for 50 000 consecutive generations. The tree with the highest likelihood was retained and is presented here (Fig. 6). Bootstrap support values (BS) were estimated from 1000 replicates using the same model in Garli. Analytical runs were performed on the CIPRES Science Gateway (Miller et al. 2010). All trees were drawn with the aid of FigTree ver. 1.3.1 (Rambaut 2009). Afrosyrphus Curran, 1927 Type species Curran, 1927, by monotypy.

Differential diagnosis (adapted from Vockeroth 1969)
Moderately robust, rather pilose species with extremely long, slender antennae, obscure abdominal markings and densely pilose hind legs. Face almost straight in profile, with low tubercle, densely pale pruinose laterally, with a shining, broad, dark median facial vitta. Eye apparently bare (with extremely short and scattered pile), holoptic in male. Antenna porrect or nearly so, with postpedicel longer than scape and pedicel together; arista bare, subbasal. Mesonotum dull, heavily pruinose; pleura dull, pruinose. Thoracic pile pale, unusually long but not obscuring ground colour. Subscutellar fringe very long and dense. Dorsal and ventral katepisternal pile patches narrowly joined posteriorly, clearly separated anteriorly. Metasternum bare. Hind coxa and hind trochanter with a pile tuft at median angle. Wing membrane with extensive bare areas on rather more than basal half, microtrichia on rest of membrane very fine but moderately dense. Ventral calypter with many fine, erect pale pile on posterior part of dorsal surface. Apical half of hind femur and hind tibia of both sexes with dense dorsal and ventral fringes of long, mostly dark pile. Abdomen unmargined, narrowly oval, long pilose.

Distribution
Afrotropical Region (see Fig. 2). Vockeroth (1969) suggested a close relationship with Epistrophe based on adult morphology and male genitalia. Láska et al. (2000) described the puparium of Afrosyrphus based on the material from Kenya collected by Schmutterer and argued that Afrosyrphus is more closely related to Epistrophella based on the puparium morphology. Our unpublished target enrichment Syrphinae phylogeny supports both of these hypotheses, with Epistrophella sister to Afrosyrphus + Epistrophe.

Differential diagnosis
This species can be distinguished from A. varipes as stated in the identification key. Overall a more robust species with slightly broader abdomen, darker pilosity in calypter, scutum and abdominal segments, and male frontal triangle with black pruinosity along eye margin that looks entirely shiny . Male Curran, 1927 (red triangles) and Afrosyrphus schmuttereri sp. nov. (black circles).

Fig. 2. Known distribution of Afrosyrphus varipes
genitalia as in Fig. 5F-H. Females are also darker than in A. varipes. In females, tergite 2 has a posterior dark fascia (tergite 2 entirely orange in A. varipes; Fig. 4A-B), the femora are darker than the tibiae (femora only slightly darker than tibiae in A. varipes) and the face is dark in background colour (orange in A. varipes). Both sexes have similar hind legs. The hind first tarsomere (= metabasitarsomere) appears orange and shining, as it has no long, black pile (Fig. 5B).

Etymology
This new species is named after its first collector, Prof. Heinrich Schmutterer, in his honour, for his dedicated work on entomology in Kenya. This species epithet is to be treated as a noun in the genitive case.

Male
Head (Figs 3D, 4C). Face with distinct, low facial tubercle and with 3-4 grooves, black medially and yellow laterally, yellow pilose with seldom black pile on dorsal half and a shiny, bare median vitta, with dense pale pruinosity covering yellow lateral areas; gena yellow, with a small black macula at eye margin, yellow pilose with pale pruinosity dorsally; lunule black, with violet iridescence; frons black, with long black pile, shiny medially, with black pruinosity along eye margin (following pale pruinosity of face) and dorsally at eye angle; vertical triangle black, black pruinose, black pilose; antennal base inflated, protruded; antennal pits clearly separated; antenna elongated, black pilose, scape partly brown, pedicel and postpedicel black; postpedicel longer than scape and pedicel together; arista black, bare; eye almost bare, with a few scattered pile; occiput pale pruinose. Thorax . Scutum black, densely pale pruinose, with long yellow and brown pile; postpronotum paler, bare, densely pale pruinose; postalar callus yellow; scutellum yellow, pale pruinose, mostly dark pilose with some pale pile on anterior and posterior margins, subscutellar fringe with long, yellow pile. Pleuron black except katatergum yellow, densely pale pruinose, yellow pilose; metaepisternum bare; metasternum bare; postmetacoxal bridge incomplete; calypter yellow, mostly yellow pilose, ventral calypter with brown fringe and fine erect pile on posterior part of dorsal surface; plumule long, pale; halter pedicel and capitulum yellow; posterior spiracular fringes yellow. Wing. Hyaline, stigma yellow except dark brown basally; membrane bare basally, cells c, br and bm entirely bare, apical cells microtrichose but bare very basally; alula microtrichose.
Legs. Coxae and trochanters black, mostly yellow pilose; fore and mid femur yellow at base very narrowly and on apical half, darker basally; fore femur with long pile posteriorly, black apically and yellow basally; mid femur similar, but also with long pile anteriorly; fore and mid tibia yellow, yellow pilose; fore and mid tarsomeres yellow; hind coxa and hind trochanter with tuft of black pile medially; hind femur and hind tibia brown; hind femur with dense dorsal and ventral fringes of long pile, yellow on basal half and black on apical half; hind tibia with dense dorsal and ventral fringes of long black pile; hind tarsomeres pale.
Abdomen (Fig. 4B-C). Unmargined, narrowly oval, long pilose. Tergite 1 yellow, densely pale pruinose, yellow pilose; tergite 2 yellow on basal half and posterior margin, black on posterior half, with posterior margin yellow, pale pilose anteriorly and laterally, dark pilose medially on posterior half, densely pale pruinose; tergite 3 yellow on basal half, black on posterior half and on lateral margins, yellow pilose laterally and dark pilose medially, with some pale pile on anterior margin, densely pale pruinose, with two fasciate areas on posterior half less pruinose; tergite 4 black with a yellow fascia on posterior margin, with two medial yellow maculae on anterior half, yellow pilose laterally and dark pilose medially, densely pale pruinose, with two fasciate areas on posterior half less pruinose, smaller than on tergite 3; tergite 5 black, pruinose, black pilose. Sternite 1 yellow, with long yellow pile; sternites 2 and 3 yellow anteriorly and black posteriorly, with long yellow and black pile; sternite 4 black, long black pilose.

Female
Similar to male except normal sexual dimorphism and as specified in diagnosis.

Variation
Based on studied material, colouration of pile on scutum and scutellum may vary slightly in amount of dark pile, ranging from little to many.

Biology
Schmutterer used larvae of Afrosyrphus to study their biological response as predators of common aphid species in East Africa (Schmutterer 1972b) or as prey for East African ants (Schmutterer 1972a). Schmutterer (1974) reared larvae of Afrosyrphus feeding on several hosts on six species of plants belonging to five different families from Kenya (preys and host plants are summarized in Table 1). He wrote that larvae of Afrosyrphus "of older stages of development stand out in comparison to those of many other aphidophagous Syrphidae of East Africa by their relatively broad, strongly flattened body. The greenish yellow of the last larval stage turns brownish 2-3 days before pupation. The drop-shaped, brown pupa has at its rear end a relatively long extension formed from a pair of stigmatic tubes."  Schmutterer (1974) also stated that "in laboratory experiments, mature larvae ready to pupate are quiescent under certain conditions and can survive several weeks without damage". He noted that at constant temperature, low humidity causes the quiescence, and guessed that Afrosyrphus spans longer periods of drought as a quiescent larva under field conditions. This could be the reason why adults almost disappear in the highlands of Kenya during dry seasons (Schmutterer 1974). Ssymank (2012) collected four adult specimens of A. varipes between 10:00 and 12:00 at full sun on a hilltop with a single flowering tree, Phyllanthus discoideus (Baill.) Müll.Arg. (Euphorbiaceae). However, specimens of the new species from Taita Hills were collected between 10:00 and 11:00, hovering high between large trees in the shade. These specimens were all males and clearly using the opening between the trees on the slope as a landmark while waiting for females to appear. This landmark mating strategy is common in syrphids and similar to hilltopping, but the flies may use any type of landmark from a trail to a forest opening, rocks or a particular tree (Skevington 2008). The trees in the shade were close to the forest margin of a primary lower montane cloud forest (Fig. 1A), in a small valley invaded by the South American invasive angel's trumpet of the family Solanaceae, Brugmansia suaveolens (Humb. & Bonpl. ex Willd.) Bercht. & J.Presl. Schmutterer (1974) reported larvae of Afrosyrphus feeding on the aphid Myzus persicae (Sulzer, 1776) from the closely related invasive Solanum seaforthianum Andrews; thus, B. suaveolens might be the host plant of the aphids for the Afrosyrphus larvae in the Taita Hills area.
Regarding the specimens collected from Bwindi Impenetrable National Park (Uganda), they were collected between 10:00 and 11:00, hovering in a sunspot in dense, moist forest over the top of a small tree approximately 3 m high, at heights of between 3 and 6 m (Fig. 1B). Once the area fell into shade, no more specimens could be found.
The type locality is in Nairobi City, probably the urban park and Nairobi riverside in or near Chiromo. According to the host plants in Table 1 reported by Schmutterer (1974), the vegetation and habitat is ruderal, with tall herb vegetation on deep soils probably adjacent to the Nairobi River. The larvae obviously use different aphid species on a variety of plant families, including aphids living on up to 4 m tall herbs like the spinach rhubarb (Rumex abyssinicus Jacq.) and the invasive Brazilian nightshade (Solanum seaforthianum). Some humidity may play an important role in supporting good aphid populations and the larval development of Afrosyrphus. We assume, based on our own records, that the Table 1. Reported prey, host plants and localities for Afrosyrphus schmuttereri sp. nov. in Kenya (adapted from Schmutterer 1974). Hemipteran names follow Favret (2019) and plant names follow The Plant List (2013).

Distribution
Species known from Uganda and Kenya (Fig. 2).

Differential diagnosis
This species can be distinguished from A. schmuttereri sp. nov. as stated in the identification key.
Overall a smaller species with slightly narrower abdomen, paler pilosity on calypter fringe, scutum and abdominal segments, and male frontal triangle with pale pruinosity along eye margin (Fig. 4D). Male genitalia as in Fig. 5C-E, with ventrally pointed surstyli and postgonites with a large triangular posterodorsal process (see also Vockeroth 1969: fig. 29). Females are also lighter than in A. schmuttereri sp. nov. Tergite 2 is entirely orange in A. varipes (with posterior black fascia in A. schmuttereri sp. nov.), the femora are only slightly darker than the tibiae (femora darker than tibiae in A. schmuttereri sp. nov.), and the face is pale in background colour (black in A. schmuttereri sp. nov.). Both sexes have similar hind legs. The hind first tarsomere (= metabasitarsomere) appears dark and bristly due to the presence of long, black pile (Fig. 5A).

Discussion
The African hover fly fauna remains one of the most poorly collected and studied continental faunas in the world. Finding a new species in a previously monotypic genus is a significant discovery that helps us better understand the morphological boundaries of the lineage. The description of a new species here also reassigns known ecological and larval information previously attributed to A. varipes. Further study on A. varipes is now warranted, as we can only assume that the larvae perform similar functions. Recent increased interest in the African fauna will undoubtedly turn up many other important new discoveries. From the currently known geographical distribution and altitudinal range (from ca 1011 to 2314 m a.s.l.) we can assume that Afrosyrphus schmuttereri sp. nov. is essentially an Eastern arc Afromontane species (in the Afromontane-Afroalpine Biotic Zone; Happold & Lock 2013), which may potentially occur in and around all lower and higher cloud forest relics, and possibly even southwards up to the Drakensberg in South Africa, although we do not have data. Afrosyrphus varipes, on the other hand, seems to be linked to the Rainforest Biotic Zone in the western and central part of the African continent, with a much lower known altitudinal range (from 400 to 1265 m a.s.l.).
A serious threat to the primary habitat of Afrosyphus schmuttereri sp. nov. is the rapid loss of natural primary cloud forests in the region (Rogers et al. 2008;Wagura 2014) due to overuse of water, as well as spreading corn fields and settlements into the cloud forest areas. We believe that the spraying of pesticides may be an important threat as well, as some of the reported host plants are cultivated as edible plants or for medical use (Rumex abyssinicus Jacq. and also Baccharoides lasiopus (O.Hoffm.) H.Rob.; see Useful Tropical Plants Database 2019). Consequently, as important pollinators and flower visitors, hover flies should be protected as far as possible by the use of selective pesticides when chemical control measures are applied against homopterous insects. Nevertheless, some agricultural use and anthropogenic changes can promote secondary urban or riverine vegetation, which is obviously suitable for Afrosyrphus larvae. The loss of larger areas of the natural cloud forests will inevitably also change the water balance and reduce the available secondary habitats, as brooks and small rivers will dry out and no longer have lush herbaceous vegetation.