Gondwanamyia, a new empidoid (Diptera) genus of uncertain placement

Abstract A new minute-size empidoid fly genus, Gondwanamyia gen. n. and two new species (Gondwanamyia chilensis Cumming & Saigusa, sp. n., Gondwanamyia zealandica Sinclair & Brooks, sp. n.) are described, illustrated, and their distributions mapped. The family and subfamily assignments remain uncertain, but features of the female terminalia potentially suggest Trichopezinae (Brachystomatidae).


Introduction
The higher classification of the exceedingly diverse Empidoidea, commonly referred to as dance flies and long-legged flies, consists of five well-defined families, namely Empididae, Hybotidae, Atelestidae, Brachystomatidae and Dolichopodidae s.l., plus three small unassigned genus-groups (Sinclair and Cumming 2006) that are sometimes given separate family status (i.e., Homalocnemidae, Oreogetonidae) (Thompson 2009;Pape et al. 2011;Marshall 2012;Sinclair, in press), or probably soon will be (i.e., Iteaphila Zetterstedt group). There are still numerous new empidoid genera awaiting description that can all be placed confidently within these various groups. However, in this paper we report on a remarkable new empidoid genus that we cannot positively assign to either family or subfamily.
The genus contains two minute-size new species, one from southern Chile and the other from the North Island of New Zealand. Here we formally describe this new genus and its two included species. The potential phylogenetic placement of this curious Southern Hemisphere group within the Empidoidea, is also discussed.

Material and methods
This study is based on material borrowed from or deposited in the following institutions: Canadian National Collection of Insects, Ottawa, Canada (CNC); Biosystematics Laboratory, Kyushu University, Fukuoka, Japan (KUMF); National Museum of Wales, Cardiff, UK (NMWC); New Zealand Arthropod Collection, Landcare, Auckland, New Zealand (NZAC). Label data for primary types are cited from the top downward, with the data from each label in quotation marks. Labels are cited in full, with original spelling, punctuation, and date, and label lines are delimited by a slash (/). The repository of each type is given in parentheses. Secondary type data are abridged and listed alphabetically.
Terms used for adult structures primarily follow Cumming and Wood (2009), except for the antenna and wing venation, where the terms of Stuckenberg (1999) and Saigusa (2006) are used, respectively. Male and female abdomens of certain specimens were macerated in hot 85% lactic acid and immersed in glycerin in order to examine terminalia. Diagnosis. Body size minute, 1.1-1.4 mm. Arista-like stylus very long, longer than thorax, lacking basal article; males and females dichoptic, eye facets convex, appearing larger than in other empidoids; clypeus strongly convex; mouthparts with large epipharyngeal carina and paired epipharyngeal blades; wing with weakened fold near base, only two longitudinal veins fully developed (R 4+5 and M 4 ), R 4+5 branched; abdomen with abdominal plaques; male terminalia symmetrical and unrotated, epandrium and hypandrium desclerotized and fused together, epandrial lobe and cercus projected anterodorsally, phallus tubular with or without an elongate apical filament, ejaculatory apodeme slender, rod-like; female terminalia sclerotized, sclerites beyond segment 6 mostly bare; tergite and sternite 8 anteroventrally articulated; female cercus sclerotized, projecting horizontally.
Wing: Length: 1.1-1.5 mm. Cuneate and narrowed at base, slightly infuscate; anal lobe and alula not developed. Costa circumambient with basal costal seta and several subequal setae proximal to R 1 ; slender, erect costal setae widely distributed to beyond R 4 ; costal break distal to Sc, continuing posteriorly as weakened transverse fold across cell bm; costa with second break at R 1 ; R 1 very short, terminating in basal 0.25 of wing; R 2+3 very short, nearly vertical, terminating at apex of R 1 ; two longitudinal veins fully developed (R 4+5 and M 4 ); R 4+5 branched, with R 4 nearly perpendicular to R 5 ; M 4 nearly straight, slightly arched toward wing margin; cell dm absent; cell bm quadrate; cell br longer than cell bm; cell cua weakly open parallel to wing margin. Halter long, subequal in length to scutum; shaft with several basal setulae; knob tapered apically.
Etymology. Named after the southern Mesozoic continent of Gondwana, in reference to the probable age and distribution of this genus of flies on the southern continents of South America and Australasia (Figs 10, 11). All five authors are responsible for the new genus name. TS first identified this new genus, and both ARP and JMC + BJS independently studied separate series. SEB recently discovered the New Zealand species. Recognition. This species is distinguished from G. zealandica by the male mid femur and tibia with stout posteroventral setae, mostly membranous epandrium + hypandrium, and long phallic filament.
Female. Similar to male, except posteroventral row of stout setae on mid femur and tibia absent. See genus description for details of female terminalia (Figs 8, 9).
Etymology. The species name refers to the country of the type locality.
Remarks. This species is confined to southern Chile (Fig. 11) where it is found in mostly damp temperate Nothofagus habitats from lowlands to tree line. All specimens were collected with yellow pan traps, except the Newton & Thayer sample, which was collected with a flight-intercept trap. Recognition. This species is distinguished from G. chilensis by the less stout posteroventral row of setae on the mid femur and tibia, presence of acrostichal setae, arista-like stylus slightly shorter, posterior setae on male sternite 8 longer and stouter, and phallus without an apical filament.
Legs: Mid femur with posteroventral row of pale setae, longer than width of tibia. Mid tibia without posteroventral row of stout setae. Apex of hind tibia slightly expanded with posteroapical comb.  Wing: Length 1.1 mm. Erect costal setae more conspicuous than in G. chilensis; subapical vein streaks present, representing veins M 1 and M 2 .
Abdomen: Four pairs of posterior setae on male sternite 8 long and stout. Male terminalia: Cercus broadly expanded at base, tapered to narrow apex; bearing fine setae. Hypandrium + epandrium with broad ventral connection; slender surstylus with pair of preapical setae; apex arched medially. Hypandrium sheath-like, with two pairs of tapered and apically pointed anterior hypandrial lobes; gonocoxal apodeme very slender, straight, arising on interior of hypandrial sheath. Phallus with hood-like apex, without apical filament; ejaculatory apodeme rod-shaped, very long and slender, running along posterior margin of hypandrium.
Female. Unknown. Etymology. The species name refers to the country of the type locality.
Remarks. This species is known only from the type-locality, on the North Island of New Zealand (Fig. 10). The New Zealand habitat is apparently very different than that for G. chilensis, but collection of additional specimens and observations on the precise micro-habitat of this species in the Matuku Reserve are required.

Discussion
Members of Gondwanamyia are among the smallest empidoid flies known. Based on the structure of their mouthparts, with free epipharyngeal blades, it is probable that they are predators on tiny arthropods, but nothing is currently known of their habits. Collection dates from both Chile (Dec., Feb., April, June, July) and New Zealand (April) indicate that this is primarily a late flying autumnal genus, apart from the December record. The weather in Chile and New Zealand during these months can be cool to cold. Most specimens were collected using yellow pan traps on the ground, indicating that members of this genus possibly fly low in the forest understory.
The small size of Gondwanamyia has presumably led to reductions in certain features such as wing venation. This has undoubtedly resulted in a loss of informative synapomorphies resulting in some confusion with homologies, creating uncertainty with the family and subfamily assignment of the genus. For example, the homologies of the male terminalia of both species are not completely resolved. The epandrium and hypandrium of G. zealandica appear fused together (Fig. 7), as presumably they are in G. chilensis, although with more desclerotization (Fig 6). In addition, the hypandrial lobe in G. zealandica has a lateral and medial component that could include the postgonite, whereas the similarly positioned large undivided lobe in G. chilensis may also incorporate the surstylus.
Gondwanamyia obviously does not belong to the Hybotidae or Dolichopodidae s.l., based on presence of a branched R 4+5 wing vein, unrotated male terminalia, and absence of a palpifer and fore tibial gland. It cannot be placed in the Atelestidae based on a branched R 4+5 wing vein, setose scape, elongate hypandrium, and presence of female tergite 10. The genus is also clearly not assignable to either the Homalocnemis Philippi, Oreogeton Schiner, or Iteaphila genus (or family) groups.
Gondwanamyia might represent a new lineage of Empididae, although no convincing synapomorphies appear to align it with any included member of that family. The male terminalia of the new genus show some similarities to most Nearctic species of Chelipoda Macquart (Empididae: Hemerodromiinae) (MacDonald 1993, fig. 12), particularly in the fusion of the epandrium and hypandrium. However, no other features of the terminalia are shared between these two groups and not all species of Chelipoda exhibit the epandrial-hypandrial fusion (Plant 2007). Although wing venation reduction is also common among certain genera of Hemerodromiinae (Plant 1993(Plant , 2005, the pattern of reduction in both the Hemerodromiinae and in Gondwanamyia does not appear to be homologous. Males of the monotypic New Zealand hemerodromiine genus Monodromia Collin, also have a wing with a similar weakened fold near the base. The function of this weakening (present in both sexes in Gondwanamyia) is unknown in both genera.
Th e form of the female terminalia of Gondwanamyia (only known for G. chilensis) suggests inclusion in the Brachystomatidae based on the anterior articulation of tergite and sternite 8, the more thickly sclerotized terminal segments with absence of micro-setae, and well developed setae at the apex of the cerci and tergite 10 (Sinclair and Cumming 2006). If the genus does belong in this family, it is not assignable to the Ceratomerinae due to the absence of a conus on the antennal pedicel, or to the Brachystomatinae due to its shortened cell cua, absence of a lacinia and straight spermatheca (Sinclair and Cumming 2006). This implies that Gondwanamyia by default, probably belongs in the Trichopezinae. This subfamily is a very heterogenous group and many genera (including Gondwanamyia) lack the internal median apodeme that projects anteriorly from female tergite 8, which originally formed the basis for recognizing this lineage (Sinclair 1995). Further study and analysis of the Trichopezinae is required to determine more conclusively whether Gondwanamyia should be assigned to this lineage.