Natalimyza milleri, Barraclough & McAlpine, 2006

Barraclough, David A. & McAlpine, David K., 2006, Natalimyzidae, a new African family of acalyptrate flies (Diptera: Schizophora: Sciomyzoidea), African Invertebrates 47, pp. 117-134 : 124-133

Natalimyza milleri sp. n.

Figs 1–12

Etymology: The species name recognises the contribution of Dr Ray Miller to our understanding of the Natalimyzidae . He was involved in collecting the entire type series.

Diagnosis:

External characters: Head usually at least partly dark; head and eye more or less rounded; frons not excavate; cheek strongly receded anteriorly; differentiated vibrissa lacking, but series of longer, anteroventrally directed cheek bristles present; medial proclinate fronto-orbital present, proclinate setula present anterior to lateral proclinate fronto-orbital towards antennal base; facial region relatively broad, width reaching length of third antennal segment in profile; palpus with apical, outstanding bristle; three (rarely four) well developed sternopleurals; four (1+3) dorsocentrals; posterior intra-alar present; costa extending to apex of vein R 4+5; crossveins r–m and m–cu well separated, distance 1.0–1.7 times length of m–cu.

Male postabdomen: Distinguished from sibling species with a similar talon-like cercus bearing two apical spines (see Figs 7, 8 ) as follows: epandrium with margin unmodified, and lacking outstanding setulae; surstylus ( Fig. 8 ) relatively broad at base, apicoventral section broadly concave such that apical half is substantially narrowed and apex acutely rounded; hypandrium ( Fig. 9 ) with anterior section bearing two pairs of relatively weakly developed spines, these slightly longer than cercal spines, and arising directly from surface (i.e. not from protruding cuticle), typically medioventrally directed such that they oppose each other.

Description:

Dimensions (mm): Male: Body length 2.7–3.0; wing length 2.5–2.8. Female: Body length 3.1–3.6; wing length 2.8–3.5.

Colour/Pruinescence: Head (including antennae) largely yellow to yellow-brown, cheek entirely so; fronto-orbital plate, ocellar triangle, part of occiput and much of arista usually a darker brown which is overlain with silvery pruinescence. Thorax yellowbrown to brown, anterior margin of pleuron usually yellowish; pruinescence relatively dense and silver. Legs, including all tarsi, entirely yellow, noticeable pruinescence lacking. Haltere and wing veins yellow.

Head ( Figs 2, 3 ): Somewhat rounded, not flattened in profile; eye of similar length and depth, ommatrichia not or barely evident. Cheek strongly receded anteriorly in a line anterior to medial proclinate fronto-orbital in profile; differentiated vibrissa lacking, a small anteriorly directed bristle at anterior extent of cheek (rarely duplicated), followed by series of usually three to four, but occasionally five to six, much longer bristles, these positioned towards ventral margin and anteroventrally directed; cheek otherwise often with a few anteroventrally directed setulae, which are more or less continuous with several rows of setulae on lateral part of occiput. Occiput prominent in profile; setulae on lateral margin downwardly directed, one or two rows of upwardly directed setulae medial to these, positioned from about mid-height of occiput to just behind and below verticals. Dorsal head bristles well developed; inner vertical the longest, usually markedly so. Fronto-orbital bristles: medial proclinate present; one proclinate setula present anterior to lateral proclinate fronto-orbital towards antennal base. Frons not excavate, entirely visible in profile; ocellar triangle usually just visible, with very few, short setulae. Fronto-orbital plate with scattered setulae. Antenna inserted noticeably above mid-height of eye; first segment exposed in profile. Facial region quite strongly recessed, relatively broad, width reaching length of third antennal segment in profile. Palpus with apical, outstanding bristle.

Thorax: Three sternopleurals, but occasionally an additional anterior bristle; anterior bristle the shortest, and the posterior the longest. Four well developed dorsocentrals (one presutural and three postsutural); prescutellar pair reaching almost twice length of either of the anterior two pairs. Posterior intra-alar present. Prescutellar acrostichals absent. Mesoscutal setulae irregularly distributed, never in distinct rows. Basal scutellars one-third to one-half length of apicals. Legs: hind basitarsus slightly to noticeably longer than mid basitarsus.Wing ( Fig. 4 ): dorsal costagial bristle 1.0–1.5 times length of ventral bristle. Costa extending to apex of vein R 4+5 or just beyond, macrotrichia in basal half of wing usually shortest opposite subcostal cell and longest between insertion of subcosta and apex of vein R 1; distance on margin between veins R 2+3 and R 4+5 one-third to one half that between veins R 4+5 and M 1. Crossveins r–m and m–cu relatively well separated, distance between them 1.0–1.7 times length of m–cu; m–cu is positioned at basal onethird to three-sevenths of wing.

Male postabdomen ( Figs 5–9 ): Protandrial sclerites ( Figs 5, 6 ): Tergite 6 relatively broad and well developed, length about half to two-thirds that of tergite 5; setulae distributed mainly near right posterior margin. Sternites 6 and 7 with setulae always lacking on about anterior one-third. A slender vestige (probably of tergite 7), evident at right hand side, lacking setulae. Sternite 8 with setulae over much of surface. Spiracles positioned as in Fig. 6 . Terminalia ( Figs 7–9 ): Epandrium unremarkable, marginal projections or other modification lacking; setulae irregularly distributed, outstanding setulae or bristle(s) absent. Surstylus prominent in profile, relatively broad at base, but with apicoventral section broadly concave such that apical half is narrowed; apex acutely rounded; a cluster of setulae often present near inner basoventral margin, sometimes a few short scattered setulae on outer, apical one-third to two-fifths. Cercus arising near base of surstylus in profile and largely exposed above it, although part of apical onequarter concealed by surstylus apex; shape unusual and complex, broadest at base, but more distally resembling a downwardly directed talon with apical region posteriorly directed and bearing two conspicuous dark, slender, apical spines; setulae relatively elongate and backwardly directed near posterior margin, thereafter much shorter and largely confined to inner margin as a medially directed series along much of length. Hypandrium U-shaped, presenting as paired sclerites united anteriorly; laterally with a posteriorly directed, short, bluntly tapered lobe on each side; paired, usually medioventrally directed, sharply pointed, black spines present on each side near anterior end (occasionally a small supernumerary present), these not arising from protrusions on cuticle, slightly to noticeably longer than cercal spines and one typically slightly shorter than the other. Gonites slender and relatively sharply pointed apically, upcurved posteriorly in profile, closely flanking basiphallus. Aedeagal apodeme very short, posteriorly directed, length at most twice that of hypandrial spines.

Female postabdomen ( Figs 10–12 ): Tergite 6 with short bristles differentiated along posterior margin, and with a few short setulae distributed just anterior to this. Tergite 7 well developed, posterior margin virtually straight, vestiture similar to Tergite 6. Tergite 8 relatively well sclerotised, short setulae present near posterior margin. Cerci posteriorly directed, broader basally, length slightly more than maximum width. Sternites 6 and 7 with irregularly distributed setulae over posterior one-third. Sternite 8 with profuse, short vestiture over much of surface. Epiproct evident as an inconspicuous transverse vestige between cercal bases; with a few medial setulae. Hypoproct small, inconspicuous. Spermathecae relatively large, subspheroid, duct gradually widened apically and not inflexed into base of vesicle; surface of vesicle inconspicuously punctate, especially on extremity opposite entry of duct.

Holotype: SOUTH AFRICA: KwaZulu-Natal: ơ ‘S. AFRICA: Natal / Van Reenen / 28 ° 22'S: 29 ° 23'E 1940m / 21–25. iii. 1984 / P. Stabbins & R. Miller // HOLOTYPE ơ / Natalimyza milleri / Barraclough & McAlpine’ [second label with red circumference]. In good condition.

Paratypes:8ơ 37^with same data as holotype; 2ơ 4^Van Reenen’s Pass, Windy Corner, 2829Ad, 25.iii.1984, alt. 1680 m, Raymond M. Miller & Patricia A. Stabbins; 3ơ 1^Giant’s Castle Game Reserve, 2929Ab, 20– 21.ii.1982, alt. 1900 m, grassland swale/along river, Raymond M. Miller.

Other material: 1ơ 1^Van Reenen’s Pass, 28 ° 22'S: 29 ° 23'E, 21–25.iii.1984, alt. 1940 m, P. Stabbins & R. Miller .

Discussion: We have seen at least four additional species from Van Reenen’s Pass and Giant’s Castle, three of which occur through late summer into autumn (late February to late April). The fourth is a winter species (late June), which is rapidly distinguished from all the others by having the head flattened in profile (see discussion under generic description). Two of the first-mentioned three species have talon-like male cerci comparable to those of N. milleri , but otherwise are readily distinguished (e.g. with distinctively shaped surstyli, differently developed and/or positioned hypandrial spines, and the apical part of the cerci not cranially directed).

There is variation in the distribution of vestiture on the surstylus, but we consider this to be intraspecific variation. There is also variation in body colouring, particularly of the thorax, which ranges from a pallid yellow-brown to a distinctly darker brown. Males of both colours have been dissected, and no discernible differences in the male terminalia were detected. A more detailed review of the KwaZulu-Natal fauna of Natalimyzidae may reveal specimens of N. milleri from additional localities, although the species is probably restricted to the Drakensberg.

RELATIONSHIPS

In considering the possibility of relationships between Natalimyza and other schizophoran superfamilies, it seems safe to eliminate first those superfamilies which disagree with Natalimyza in the uniform absence of vein 7 in that part of the wing lying beyond the alula, especially as there appear to be no striking morphological similarities to any of their component taxa. These are: Diopsoidea (Nothyboidea), Opomyzoidea (Agromyzoidea), Asteioidea, and the Ephydroidea (Drosophiloidea). The putative exceptions to the absence of vein 7 in the ephydroid families Curtonotidae and Cryptochetidae have been carefully investigated and found baseless. The natalimyzid combination of convergent postvertical bristles, distinct preapical dorsal tibial bristles, unbroken costa, and the characteristic arrangement of the protandrial sclerites gives an immediate additional means of separation from representatives of all such superfamilies. Those schizophoran superfamilies which show evidence of the presence of the distal section of vein 7 in their groundplans are: Conopoidea (as delimited by Colless & D.K. McAlpine 1970); Sciomyzoidea (D.K. McAlpine 1991 a, i.e. Sciomyzoidea + Lauxanioidea of Hennig 1973 and J.F. McAlpine 1989); Heteromyzoidea 1; Nerioidea (or Micropezoidea); Chloropoidea (see D.K. McAlpine 1982); Tephritoidea (see J.F. McAlpine 1989); Muscoidea (see Hennig 1958). Natalimyza cannot belong in the Conopoidea because it lacks the characteristic elongate proboscis with attenuated labella, it has a normal notopleural region, it has not lost the mid-coxal prong, veins R 4+5 and M 1 are not convergent, the anal cell is not long and pointed, and the male postabdomen is not symmetrical.

Natalimyza is less strongly differentiated from the Heteromyzoidea (= Heleomyzoidea), particularly as it has the combination of convergent postvertical bristles, often more or less differentiated vibrissa, and distinct preapical dorsal tibial bristles, which, though rather characteristic of Heteromyzoidea , occurs elsewhere (e.g. some

1 The superfamily Heteromyzoidea [synonyms being the Heleomyzoidea as delimited by Colless & D.K. McAlpine (1970) and the Sphaeroceroidea of J.F. McAlpine (1989)] and the family Heteromyzidae (including Heleomyzidae s.l. and the Sphaeroceridae of recent authors) are used in the revised sense of D.K. McAlpine (MS submitted for publication). This is the simplest means of combining phylogenetic consistency with nomenclatural priority under the International Code of Zoological Nomenclature.

Drosophilidae , very few Lauxaniidae ). Unlike Natalimyza , the Heteromyzoidea have almost invariably the costa broken at the subcostal position, and the mid basitarsus at least slightly more elongate than the hind basitarsus. Some chyromyids, among forms sometimes placed in the Heteromyzoidea , have lost the costal break, but these are reduced types of minute size and do not otherwise approach Natalimyza . A rapid comparison of Natalimyza and the Chyromyidae gives a further 10 apparent groundplan differences between the two groups, which need not be detailed here.

The only heteromyzoids with a visible distal section of vein 7 are found in certain taxa of the family Heteromyzidae (or Heleomyzidae s.l.).Almost throughout this family a well sclerotised distal section of vein 6 is present and, in those taxa with vein 7 visible beyond the alula, vein 6 is long, reaching the margin or almost so; reduction of veins in this part of the wing always entails loss of the distal section of vein 7 before any notable reduction of vein 6. Thus Natalimyza presents a reduction pattern in the anal veins which is unlike any of the varied types currently included in Heteromyzoidea . Taking this together with the above anomalous character states, we feel it to be improbable that Natalimyza belongs in the Heteromyzoidea .

The phylogeny and relationships of the Nerioidea have been under study (see particularly D.K. McAlpine 1996). The postabdominal morphology of both males and females of Natalimyza cannot be derived from that of the groundplan of the Nerioidea. Also, typical preapical dorsal tibial bristles, like those present in Natalimyza , are absent in the Nerioidea, though there are several scattered dorsal tibial bristles present in some nerioid taxa. No other special points of resemblance have been found, and the Nerioidea can reasonably be eliminated from the taxa likely to be closely related to Natalimyza .

The Chloropoidea (sensu D.K. McAlpine 1982) includes taxa, such as certain of the Milichiidae and Canacidae , with vein 6 very curtailed beyond the anal cell and vein 7 visibly developed as in Natalimyza . However, the following character combination in Natalimyza indicates that any close relationship to the chloropoid families is unlikely: costa unbroken; subcostal cell sclerotised distally; each tibia with one distinct preapical dorsal bristle; prosternum quite narrow. The families of Chloropoidea have received different treatment by Brake (2000) and Buck (2006), and we are not entirely in agreement with some of their reasoning and conclusions. The points of disagreement have little relevance to the present discussion.

The superfamily Tephritoidea is most significantly characterised by the modifications of the female postabdomen (J.F. McAlpine 1989). Natalimyza lacks these modifications and retains a much more plesiomorphic female postabdomen. In the absence of any other notable points of resemblance, Natalimyza can be safely excluded from the Tephritoidea .

The Muscoidea (sensu Hennig 1958 or Calyptratae of J.F. McAlpine 1989) have been apportioned a large number of groundplan character states by J.F. McAlpine. Evaluation of many of these will depend on future investigation, but at least the following appear to have diagnostic value at present: frons with continuous sclerotised frontoorbital plates connecting vertex to parafacial and bearing a series of incurved bristles anteriorly; antennal segment 2 with dorsal longitudinal seam or slit; mesoscutum with transverse suture of each side prolonged medially as a groove; preabdominal spiracles situated in tergites; sternite 5 of male bifid. As Natalimyza has none of these features, it is certainly excluded from the Muscoidea.

Natalimyza possesses a character state combination which is more readily reconcilable with the Sciomyzoidea than with other schizophoran superfamilies. The following natalimyzid features are particularly relevant: incurved lower fronto-orbital bristles absent; antennal segment 2 without dorsal slit; preapical dorsal bristle present on all tibiae; lower calypter not forming a lobe; costa unbroken; subcosta complete distally, diverging from vein R 1; anal and second basal cells complete; vein 7 discernible well beyond alula; male postabdomen with abdominal sternites 6 and 7 displaced towards left side, with aedeagus short, not coiled; female postabdomen without any postabdominal segments modified to form an ovipositor or ovipositor sheath, with cerci separate.

Natalimyza differs from most Sciomyzoidea in having the mid and hind basitarsi similarly slender and the former usually slightly to noticeably shorter than the latter. Typical sciomyzoid families (like heteromyzoid families) have the mid basitarsus more elongate (either longer or more slender or both) than the hind basitarsus, but this feature is variable in some sciomyzoid families.

The points of agreement with the Sciomyzoidea may not provide very strong evidence for relationship between Natalimyza and the other families classed in Sciomyzoidea , because the above character states are either apparently plesiomorphic for the Schizophora or occur in representatives of some other schizophoran superfamilies. However, this combination of character states does not occur in any other superfamily. Also the limits of the Sciomyzoidea are not agreed upon, and it is very difficult to define the superfamily by means of sound apomorphic character states (see also Yeates & Wiegmann 2005: 30). The sciomyzoid families are associated together largely because they span a gradation of character state combinations, which causes difficulty in demarcating narrower groupings. D.K. McAlpine (1991 a, b) discussed some of the difficulties, and adopted a broader scope for the Sciomyzoidea to include the Lauxanioidea as defined by Hennig (1958) and J.F. McAlpine (1989).

The wider superfamily Sciomyzoidea as here delimited includes the following families: Sciomyzidae (possibly including Phaeomyiidae ), Huttoninidae , Helosciomyzidae , Heterocheilidae , Dryomyzidae , Helcomyzidae , Coelopidae , Sepsidae , Ropalomeridae , Eurychoromyiidae , Chamaemyiidae (possibly including Cremifaniidae) and Lauxaniidae (including Celyphidae ). The last three families are placed in the Lauxanioidea by Hennig and J.F. McAlpine (see above).

The principal structural autapomorphy of the Sciomyzidae is the reduction of the prelabrum and its remoteness from the margin of the face.Also the Sciomyzidae do not have a well developed series of cheek (peristomial) bristles; the postvertical bristles are divergent or absent; the distal section of vein 6 is always sclerotised and usually reaches the margin, but if it is shortened (e.g. in Tetanura Fallén ), the distal section of vein 7 is quite obsolete. Natalimyza has none of these features and cannot be associated with the Sciomyzidae .

The family Huttoninidae (restricted to New Zealand) has vein 6 well sclerotised but not reaching the margin; the distal section of vein 7 obsolete; abdominal sternite 1 vestigial or absent; abdominal tergites 1 and 2 not somewhat separated by a membranous dorsal line. Natalimyza has none of these huttoninid features.

The families Heterocheilidae (D.K. McAlpine 1991 b) and Helcomyzidae are seashore inhabiting flies with the mid basitarsus markedly more elongate than the hind basitarsus, and with sclerotised prothoracic precoxal bridges, features in disagreement with Natalimyza . The Coelopidae are also seashore flies and differ from Natalimyza in the terminal ventral process of the male fore basitarsus, the characteristic sternopleural bristle, sternopleural suture, and terminal tarsal segment, and the more reduced first abdominal sternite (see also D.K. McAlpine 1991 a). The Dryomyzidae and Helosciomyzidae also have the mid basitarsus notably more elongate than the hind basitarsus, at least 3 long terminal spurs on the mid tibia, and lack a definite series of cheek bristles. The last five families ( Heterocheilidae to Helosciomyzidae ) further differ from Natalimyza in having vein 6 almost reaching the wing margin and vein R 2+3 shorter, terminating further from the end of vein R 4+5.

The families Sepsidae and Ropalomeridae differ from Natalimyza in having a sclerotised distal section of vein 6, but no discernible distal section of vein 7, also one or more bristles on the margin of the metathoracic spiracle, and the postvertical bristles, when present, divergent. The Eurychoromyiidae also have a long vein 6, but no distal section of vein 7, and, like the Helcomyzidae , Heterocheilidae , and Ropalomeridae , have sclerotised prothoracic precoxal bridges.The only described species (from Bolivia, see J.F. McAlpine 1968), is so distinct from Natalimyza that detailed comparison is unnecessary.

The Lauxaniidae , like Natalimyza , usually have convergent postvertical bristles, one preapical dorsal tibial bristle on each tibia, a variably shortened vein 6 (though with a sclerotised distal section) and visible vein 7 (veins 6 and 7 more reduced in some lauxaniids with an exceptionally narrowed wing-base). They differ from Natalimyza in having the hypofacial generally broadly visible in profile, a broad prosternum, not more than 2 sternopleural bristles, a mesopleural bristle (except in the much modified Celyphinae with reduced chaetotaxy), segment 7 of female abdomen with tergite and sternite fused into a complete ring (not yet confirmed as a groundplan condition for Lauxaniidae ), and the protandrium completely symmetrical. The last condition, in particular, indicates that Natalimyza , with its much more plesiomorphic protandrium, cannot be derived directly from the Lauxaniidae .The Tertiary fossil genera Chamaelauxania Hennig and Hemilauxania Hennig , which are referred by Hennig (1965) to the Lauxaniidae , have more than 2 sternopleural bristles, but the reference of these to Lauxaniidae is at best problematical, despite Hennig’s firm assertion. Chamaelauxania may be closer to the Chamaemyiidae and has the antenna, particularly the arista, resembling that of the fossil chamaemyiid Procremifania Hennig.

Of the sciomyzoid families, the Chamaemyiidae have the wing veins of the anal region most like those of Natalimyza , the distal section of vein 6 being much shortened, and that of vein 7 being distinctly curved but unpigmented. If Chamaelauxania preserves characters of the chamaemyiid stem group, then its resemblance to Natalimyza in the greater number of sternopleural bristles (but still no mesopleural), presence of preapical dorsal tibial bristles, and apparently at least one asymmetrically placed protandrial sternite might be significant. Natalimyza differs from living Chamaemyiidae most notably as follows (character states of Chamaemyiidae in parentheses): legs relatively long (legs shortened, probably as a groundplan apomorphy), thorax yellowish to tawny brown (thorax ashen-grey, probably as a groundplan apomorphy), preapical dorsal tibial bristles present (probably lost as a groundplan apomorphy), mid basitarsus not longer than hind basitarsus (mid basitarsus markedly longer than hind basitarsus as a groundplan plesiomorphy), segment 4 of arista desclerotised (forming a sclerotised ring as a groundplan plesiomorphy), segment 5 of arista swollen, subspherical (probably enlarged, elongate as a separate apomorphy in groundplan). The significance of these distinctions is such that Natalimyza cannot reasonably be included in the Chamaemyiidae .

In conclusion, based on the available data and the characters studied, we consider that no definite sister group relationship with any of the families in the wider Sciomyzoidea is suggested. We therefore consider it appropriate to treat provisionally the Natalimyzidae as incertae sedis within the superfamily, pending a rigorous phylogenetic study of all included families.

Figures

Fig. 1. Habitus of Natalimyza milleri, female (Van Reenen’s Pass).

Figs 2–4. Natalimyza milleri: (2, 3) head, showing detail of all vestiture: (2) profile, (3) dorsal view; (4) wing (names of wing cells italicised, b2 – second basal cell).

Figs 5, 6. Natalimyza milleri, male postabdomen: (5) protandrial sclerites, ventral view (detail of dorsal structures, including tergite 6, omitted); (6) protandrogram (vestiture omitted).

Figs 7–9. Natalimyza milleri, male terminalia: (7) dorsal view, including detail of all vestiture (left margin of perianal field omitted to reveal internal structures); (8) right lateral view, showing detail of epandrium, surstylus and cerci (arrows identify external structures); (9) oblique ventral view (detail of epandrium, cerci, surstyli and other dorsal structures omitted).

Figs 10–12. Natalimyza milleri, female postabdomen: (10, 11) segments five to eight and cerci: (10) dorsal view (hypoproct omitted), (11) profile, slightly oblique to show detail of apical structures, vestiture of sternites 5 and 6 omitted (epiproct concealed beneath tergite 8); (12) spermatheca and apical section of spermathecal duct.