Proposal of Karliella gen. n. for the Afrotropical ‘Pegomya’ sexpunctata Karl, 1935 (Diptera: Anthomyiidae), a Possible Kleptoparasite of Dung-Breeding Beetles

ABSTRACT The genus Karliella Michelsen, gen. n., is described to accommodate Pegomyia sexpunctata Karl, 1935 (syn.: Pegomyia abdominalis Emden, 1941). The peculiar morphology of Karliella makes it difficult to find a closer relationship to any particular anthomyiid genus or group of genera, which suggests that it represents the only extant member of an ancient lineage. The type species and only included species, Karliella sexpunctata (Karl, 1935), comb, n., appears to be locally abundant in dry and mesic woodland of Namibia and South Africa, but has also been found in Botswana and Zambia. The larvae probably develop in dung from large herbivorous mammals. The female has been observed laying eggs on newly formed, unburied scarab dung-balls, but whether this is customary behaviour is unknown.


INTRODUCTION
The chapter on the family Anthomyiidae that I am preparing to the forthcoming Ma nual of Afrotropical Diptera will need a key to all regional genera for the adults of both sexes. The Afrotropical anthomyiid fauna appears rather depauperate, with 66 de scribed species recorded to date (Kirk-Spriggs & Stuckenberg 2009). Practically all Afrotropical species known so far may readily be assigned to one of the following 11 genera, originally proposed for Palaearctic species: Anthomyia Meigen, 1803, Bota nophila Lioy, 1864, Calythea Schnabl & Dziedzicki, 1911, Chirosiomima Hennig, 1966, Delia Robineau-Desvoidy, 1830, Emmesomyia Malloch, 1917, Enneastigma Stein, 1916, Fucellia Robineau-Desvoidy, 1841, Lasiomma Stein, 1916, Leucophora Ro bineau-Desvoidy, 1830 and Pegomya Robineau-Desvoidy, 1830. However, a single spe cies cur rently assigned to Pegomya Robineau-Desvoidy (syn.: Pegomyia Macquart, 1835, un justified emendation) (Pont & Ackland, 1980) and re corded so far from South Africa, Na mibia, Botswana and Zambia, is a remarkable ex ception. It was first de scribed by Karl (1935) as Pegomyia sexpunctata from three males and two females from Montrose, South Africa, and subsequently by Emden (1941) as Pegomyia abdominalis from two fe males from Mbala, Zambia, at the Tanzanian border. The species was described at a time when the generic classification of anthomyiid flies was somewhat preliminary, often disregarding the several characters of the male ter minalia now considered essen tial by all students of the family. The yellowish legs and abdomen, bare anepimeron and vent rally setulose vein C are all characters that the currently recognized species shares with most species of Pegomya and surely the rea son why it was first described in that ge nus. However, Emden (1941) admitted that it might be "entitled to generic rank" be cause of the presence of an apical pv seta on the hind tibiae.
The present paper reassigns Karl's species Pegomyia sexpunctata to a new monotypi cal genus, Karliella gen. n. Morphological evidence in support of this is provided in a redescription of the male and female. The coprophagy by the larvae is further incompatible with classification in Pegomya, a genus in which the larvae known so far feed on fresh tissue from either plants or fungi.

MATERIAL AND METHODS
The specimens examined in this study belong to the National Museum, Bloemfon tein, South Africa (BMSA), National Museum of Namibia, Windhoek, Namibia (NMNW), Natural History Museum, London, UK (BMNH), Natural History Museum of Denmark, University of Copenhagen, Denmark (ZMUC), and Senckenberg Deutsches En tomologisches Institut, Eberswalde, Germany (SDEI). Photographs were taken using a Leica digital camera M205A mounted on a Leica DFC 420 stereomicroscope. Images were processed with a Leica Application Suite program. Stacking and final pro cessing were carried out with Helicon Focus and Adobe Photoshop software. Adult mor phological terminology follows Cumming and Wood (2009). Abbreviations: a -anterior, ad -anterodorsal, av -anteroventral, d -dorsal, pposterior, pd -posterodorsal, pv -posteroventral, v -ventral. TAXONOMY Genus Karliella gen. n. Type species: Pegomyia sexpunctata Karl, 1935. Etymology: The new genus is named after Otto Karl (1868Karl ( -1945, the author of the type species. Karl lived in the Pommeranian city of Stolp [now Słupsk], in the north-western part of present-day Poland. As a dipterist, he published faunistic papers on Bra chycera from Pommerania and the island of Amrum at the German North Sea coast, but further contributed to the taxonomy of his favourite groups, the present families Mus cidae, Fanniidae and Anthomyiidae. Diagnosis: The following combination of characters will assist in recognition of Karliella as being distinct from other anthomyiid genera: femora and tibiae extensively or wholly light-coloured, ochre-brown to yellow; abdomen with paired dark spots at anterior margins of tergites III-V, both laterodorsally and ventrally; arista plumose, longest rays at least ⅔ as long as width of postpedicel; upper occiput bare below postocular setae; vein C bare on the dorsal surface and setulose on the ventral surface; hind tibia with 1-4 pd setae, none of them reaching half the length of hind tarsomere 1, and with a short apical pv seta. Karliella sexpunctata (Karl, 1935), comb. n.

Description:
Size. Both sexes very variable, as indicated by a wing length range of 3.0-5.5 mm.
Male. Head (Figs 1, 2): Ground colour dark on vertex, occiput and postgenae, light ochreyellow on frons, parafacials and genae, covered in light grey to silvery grey dusting; antennae brownish black or to varying extent orange-brown to orange-yellow on scape, pedicel and basal part of postpedicel; palpi brownish black, on basal half often ochreyellow dorsally. Antennae distinctly shorter than comparatively deep face; plumose arista abruptly thickened basally; longest aristal rays ⅔ to fully as long as width of postpedicel. Proboscis short, with relatively slender haustellum and small labella; palpi slender, slightly longer than haustellum. Frons on upper part narrow, with contiguous parafrontalia, bare except for 2 or 3 pairs of frontals well below middle. Gena in profile 0.12-0.21× as deep as eye height, proportionately wider in large specimens; short genal setae arranged in single row but more abundant on aristal angle. Upper occiput without setulae below short postoculars. Thorax (Fig. 5): Ground colour wholly dark or to varying extent ochre-yellow on prothoracic parts, covered in thick light grey to olive-grey dusting, on middle third of mesonotum sometimes with a median brown vitta. Prosternum bare. Dorsocentral setae 2+3; acrostichals, except near scutellum, setulose, abundant and arranged in two close-set rows. Prealar seta about the same length as but weaker than posterior notopleural seta. Scutellum with minute hair-like setulae beneath tip. Proepisternals 2, proepimerals (1-)2. Anepisternum with anterodorsal seta; katepisternal setae 1-2+2. Anepimeron, katepimeron and metapleuron bare. Wings ( Fig. 6): Wing and calypteres with a light brown tinge; haltere ochre-yellow. Vein C with short spinules in ad and av rows, bare on dorsal surface but with short setulae in two irregular rows ventrally. Lower calypter the same size as and thus distinctly projecting behind upper calypter in lateral view.
setae; fore tibia with a weak (sometimes absent) pv seta in middle, apically with d seta flanked by short ad and pd setae. Mid femur without av setae, but with 2 short pv setae near base; mid tibia with a small ad seta on distal third, 1 pd and 1 p setae on middle third. Hind femur has, in addition to sub-basal v seta, 3 or 4 short av setae on at least distal half and a few short pv setae on middle third and apically; hind tibia with 1-3 av, 2-4 ad, 1-4 (normally 2-3) pd setae, 0-1 p seta near middle, and an apical pv seta about the same length as sub-basal v seta on hind tarsomere 1. Abdomen (Figs 7-9, 12-14): Slender, conical, subcylindrical, rather tapering, with small terminalia. Covered in thin silvery grey dusting over pale ochre-yellow ground co lour except for paired brown to blackish rounded spots laterodorsally and ventrally at anterior margins of tergites III-V, sometimes even with indications of a median dark stripe on tergites II and III; caudal parts of abdomen often darkened to a varying ex tent, especially in small specimens. Robust hind marginal setae present laterally on ter gites II and III and forming complete rows on tergites IV and V; discal setae absent. Ter gite VI short, bare, incorporating spiracles VI near lateral margins; syntergosternite VII+VIII with pair of strong setae and some setulae. Sternite I setulose; sternites II-V at fore margin each with a widely spaced pair of alpha-sensilla; sternites III and IV with a pair of strong hind marginal setae. Sternite V small, cordiform, with short, sharply tapering posterior lobes. Hypopygium (Figs 15-18): Epandrium small, rather shallow, with a loose tuft of longish setulae at lower lateral sides; cercal plate on distal ⅔ forming an acutely pointed, Figs 10, 11. Karliella sexpunctata (Karl), female abdomen, dorsal and ventral views. Scale bar = 0.5 mm.
beak-like process extending freely between surstyli and bearing three pairs of robust setae; a narrow, sclerotized bridge is formed between basal opposing parts of each cer cus and surstylus; surstyli of unusual shape, split deeply into two slender arms; a narrow, sclerotized extension from epandrium unites with dorsobasal part of surstylus; ba cilliform sclerites short, fused with ventrobasal extensions of surstyli; pregonites with two setulae, at outer basal margin articulated with central plate of hypandrium; postgonites slender and bare except for group of sensilla sub-basally; outer extension of phallapodeme detached from hind margin of hypandrial plate and from inner bases of pregonites; basiphallus with epiphallic extension and articulating with a short, sessile distiphallus without paraphallic processes or sclerotized acrophallus; ejaculatory apo deme notably small.
Legs: Femora wholly ochre-yellow, with at the most tips of mid-and hind-femora narrowly infuscated. Fore tibia with larger pv seta and small ad seta on distal third. Mid tibial ad seta more robust. Hind femur with fewer (2-3) av setae confined to distal third, and only 2 pv setae, one near middle and one apically; hind tibia with fewer submedian setae: 1 av, 2-3 ad, 1-2 pd and 0 p. Abdomen (Figs 10,11,19,20): Sternite I bare. Sternites II-VII with paired alphasensilla at anterior margins. Oviscapt (6 th and following segments) very short, barely ⅓ as long as rest of abdomen, thick, moderately tapering. Spiracles VI situated within anterolateral margins of tergite VI; spiracles VII situated in membrane immediately behind lateral hind margins of tergite VI. Tergites VI and VII fully divided mid-dorsally, tergite VIII only with a posterior notch. Short cerci and hypoproct with dense cuticular pubescence, other sclerites of oviscapt only with dusting. Three spermathecae (Fig. 21), smooth and globular, with a distinct neck.
The larvae are probably coprophagous, as judged by a published observation of a female apparently laying eggs on a scarab dung ball (Emden 1941). Dung from herbi vorous mammals dessicates quickly in warm and dry climates. Accordingly, some Di ptera have evolved the habit of making their offspring feed on dung buried by scarab beetles (see Sivinski et al. 1999, and references therein), despite the obvious challenge of sharing this resource with the beetle offspring. It is of course too early to say whether this klep toparasitic behaviour on the part of Karliella sexpunctata is obligate or facultative. Whatever the case may be, the pronounced variation observed in adult body size suggests that the larvae sometimes face serious limitations in food supply.

DISCUSSION
The new genus Karliella is established for a single species endemic to but apparently widespread in the central and southern parts of the African continent. It unquestionably belongs to the calyptrate family Anthomyiidae because of the following combination of characters: (1) head holoptic in male, dichoptic in female; (2) female with a pair of crossed interfrontal setae; (3) lower calypter well developed; (4) tiny hair-like setulae present beneath tip of scutellum; (5) vein A1 extended to wing margin as a weak fold; (6) male surstyli and cerci united by vertical sclerotized connections immediately distal to lateral cercal apodemes; (7) spiracles VI and VII present in female abdomen. Beyond that, as indicated below, morphology provides no obvious clues as to the closest relationships of this taxon.
The deeply biramous surstyli in combination with the apically projecting and very pointed cerci armed with three pairs of giant setae are diagnostic male characters not seen in other anthomyiid genera. The paired dark spots found on the ventro-flexed la teral parts of tergites III-V in both sexes also represent a unique condition among an thomyiid flies.
A narrow and flexible sclerotized bridge connects the surstyli with the epandrium in male Karliella flies. Similar articulations between the surstyli and epandrium are found in the majority of anthomyiid genera including Anthomyia, Botanophila and Delia. However, as first pointed out by Hennig (1976), some anthomyiid genera differ by having a small sclerite inserted at the articulations between the surstyli and epandrium. Among Afrotropical Anthomyiidae, this is seen in Pegomya, Emmesomyia, Calythea and Enneastigma.
Dissection of pregnant female anthomyiids of species known to breed in decomposing substrates often reveals the presence of an incubated egg in the uterus with a fully de veloped larva capable of hatching during or immediately after oviposition (pers. observ.). Because the remaining eggs are unincubated, this condition has been classified as simultaneous viviparity (Meier et al. 1999). To what extent this is a facultative or obligate habit among different species of anthomyiids is an open question, but it has evidently been attained through homoplasy in various anthomyiid lineages. The antho myiid oviscapt (postabdomen) is as regards the ground plan undoubtedly slender, well set off from and hardly shorter than the preabdomen, and has slender cerci with projecting, cylindrical apices. In Karliella, the oviscapt is rather thick, only about one-third as long as the preabdomen, and with very short cerci. Despite the apparent lar val coprophagy and notably short oviscapt, it seems that Karliella sexpunctata is an ob ligate oviparous species.
The possible kleptoparasitic habits of Karliella involve dung-storing Coleoptera as hosts. Kleptoparasitism is further known as an obligate habit in the anthomyiid genera Leucophora and Eustalomyia Kowarz, 1873, in which the females lay eggs in the borrows of solitary and presocial pollen-or flesh-storing aculeate Hymenoptera (Amiet & Volkart 1983;Meyer-Holzapfel 1986;Polidori et al. 2005, and references therein). The overall structure of the male terminalia suggests that these two genera are closely re lated to each other and to Delia. They all share slender male surstyli without the typical anthomyiid trait of a subdistal, mesal incision. Karliella differs fundamentally from these by having male surstyli that are deeply incised and biramous.
The reason for the absence of characters in support of a very close relationship between Karliella and any other known anthomyiid genus or group of genera might well be that it represents the only extant member of a relict lineage with remote ties to other extant Anthomyiidae. A comparable case is the putatively old Neotropical lineage represented by the genera Phaonantho Albuquerque, 1957, Coenosopsia Malloch, 1924and Coenosopsites Michelsen, 1996 from Dominican amber) as recognized by Michelsen (1991Michelsen ( , 1996Michelsen ( , 2000. Both relict lineages actually share in the male sex a rather tapering, subcylindrical abdomen, with the terminalia and sternite V notably small, more reminiscent of Muscidae and Fanniidae than Anthomyiidae. That similarity is probably to be ascribed to symplesiomorphy, alternatively homoplasy. Hopefully, mo lecular studies will soon throw more light on the relationships of Karliella.