A review of Madagopsina Feijen, Feijen & Feijen (Diptera, Diopsidae) with description of a new species, key to the species, and discussion of intrageneric relationships

Abstract For the recently established genus Madagopsina (Diopsidae, stalk-eyed flies), Madagopsinamakayensis Feijen, Feijen & Feijen, sp. nov. is described from Madagascar. A concise catalogue is given for the genus and an identification key is presented for its six species. The differential character states are listed for the two species groups of the genus: the Madagopsinaapollo species group and the Madagopsinaapographica species group. The intrageneric relations are discussed based on morphology, geometric morphometrics analysis of wing shape, and allometric data for eye span against body length. Each of these three procedures places the new species in the M.apollo species group with Madagopsinaparvapollina as its closest relative. New records are presented for M.apographica and M.parvapollina.


Introduction
In 2018, Feijen et al. erected Madagopsina Feijen, Feijen & Feijen and Gracilopsina Feijen, Feijen & Feijen as endemic genera for Madagascar. These new genera were placed in a Diopsidae clade with irrorated wings, named the Teleopsis genus group. In Madagopsina, two earlier described species were placed: Diopsis apollo Brunetti and Diopsis (Eurydiopsis) apographica Séguy. In addition, three new species were described for Madagopsina: M. freidbergi, M. parvapollina, and M. tschirnhausi. Shortly after the 2018 publication, a single specimen of Madagopsina was received which turned out to be an undescribed species. This species is described herein. In Madagopsina, two species groups were distinguished by Feijen et al. (2018), the Madagopsina apollo species group and the Madagopsina apographica species group. Based on morphology, allometric data and geometric morphometrics analysis of wing shape, the new species is placed in the M. apollo species group. Because of the description of the new species, the sets of character states for the two species groups need to be adapted. A concise catalogue for Madagopsina is presented, as well as a new identification key to the six species of the genus. Some new Madagopsina records are included in the catalogue. The first live photographs of M. parvapollina are presented as these high-resolution pictures nicely show differential characters for the species group. In Feijen and Feijen (2021 in press), a key to the Afrotropical genera of Diopsidae is presented along with a synopsis of the Afrotropical Diopsidae fauna, including the genus Madagopsina.

Materials and methods
The description of M. makayensis Feijen, Feijen & Feijen, sp. nov. is based on a single male specimen that was preserved in alcohol. The holotype is now pinned with the genitalia placed in a genitalia tube attached to the pin. Some additional records for Madagopsina became known via photographs placed on www.iNaturalist.org. For the rate of dimorphism D, the difference between males and females in allometric slope for eye span on body length is used in the Diopsidae (Baker and Wilkinson 2001). Details on procedures for preparing genitalia slides, and procedures for taking measurements are given in Feijen et al. (2018). For information on morphological terminology and on photographic equipment used, the reader is referred to the same source. Some changes have been made to the terminology used: the aedeagus is now referred to as phallus, while the apodeme of the surstylus is now called the apophysis. The procedures for the wing geometric morphometrics analysis are described in Feijen et al. (2018). The following institutional codens and abbreviations are used:  ;Shillito 1971: 287;Feijen 1981: 482;Feijen 1989: 63;Feijen and Feijen 2013: 182, 185.

Remarks.
A concise catalogue for the genus is given below. For details on the type series, records, and combinations to various other genera of the earlier described species can be referred to Feijen et al. (2018). Reference is now made to new records which appeared after this publication.  Feijen, Feijen & Feijen, sp. nov. can be recognised by its medium size (body length ♂ 7.3 mm), brown colour (however, due to conservation in alcohol it is likely that all the brown colours would be more yellowish in a live specimen, like in the other Madagopsina species), body mainly thinly pruinose (pollinose) with few small setulae, only katepisternum and katepimeron glossy, absence of facial teeth, medium-sized inner vertical seta (1.7 × stalk diameter), scutellar spines 2.0 × as long as scutellum, quite large apical seta (45% of scutellar spine length), incrassate fore femora with around 48 tubercles, irrorated wings with three vague crossbands including an H-shaped configuration with central and preapical crossbands, wing apex infuscated, central band slightly broader than preapical band, pale wing spots in cell r2+3 and cell m1, a vague pale spot in cell m4, abdomen clubshaped, no pruinose spots on tergites, ♂ spiracles 7 in slit of synsternite 7+8, surstyli rounded and bulbous with an apically rounded apophysis, microtrichia on posterior apical third, phallapodeme with ratio posterior arm/anterior arm 1.05, straight ejaculatory apodeme with only a slight sickle-shape apically, phallus remarkably broad and sclerotised, assumed moderate sexual dimorphism with regards to eye span (D ≈ 1.0), Description. Measurements. Body length ♂ 7.3 mm; eye span 8.8 mm; wing length 5.9 mm; length of scutellar spine 1.01 mm.
Head. Central part brown, ocellar tubercle and arcuate groove dark brown; central head thinly pruinose (Figs 3-6); an elongate bulbous medial ridge in front of ocellar tubercle, parallel grooves on both sides of this ridge, lateral areas of frons flat; medial occiput flat; face convex in profile, facial corners square, no facial teeth (Figs 5, 6); clypeus small, not protruding; arista finely microtrichose on less than basal half; the rate of dimorphism cannot be calculated, but in the graph ( Fig. 31) with the allometric lines for the three species of the M. apollo species group it can be seen that the single data point is located in line with the allometric line for M. parvapollina males, while given that the slopes for the females for the species must be almost identical, it follows that D for the new species must be almost identical to the D = 0.98 for M. parvapollina or slightly higher (see also the section "Allometric aspects with regard to  this is a dimorphic species with a moderate rate of dimorphism D ≈ 1.0; stalks thinly pruinose, brown, broad apical parts dark, funiculus brown, pruinose; inner vertical seta medium-sized, 1.7 × diameter of eye stalk (Figs 3, 5), base of inner vertical seta a minor elevation, one-eighth diameter of the stalk; outer vertical seta broken off; central head and stalks with a few tiny white setulae.
Wing. Irrorated with a rather vague, brownish, H-shaped configuration; apex (8% of wing length) with brownish infuscation (convex on proximal side); 3 crossbands, the basal and central band hardly separated, a pale preapical band and three pale spots (Figs 3, 10); preapical crossband (distal leg of H) broad, marginally darker than other bands and with slightly irregular edges; preapical band connected to central band in cell r1, in cell r4+5 and around veins R2+3 and R4+5; central band slightly broader  (Feijen et al. 2018, figures 187, 184, 186, 188). than preapical band and with ill-defined proximal edge, only in cell m4 vaguely separated from basal band; basal band running from cell c to posterior wing margin, widening posteriorly; this infuscation pattern creates a pale (but not hyaline) preapical band between dark preapical band and infuscated wing apex, two pale spots between central and dark preapical bands (one in cell r2+3 and one in cell m1), and one vague pale spot centrally in cell m4 between basal and central band (Figs 3, 10); glabrous basal areas include basal apices of cell c and cell r1, basal half of cell br, basal quarter of cell bm+dm except for posterior margin and basal third of cell cua; vein M4 reaching to just beyond halfway the wing margin.
Postabdomen male. Sternite 4 a rectangular plate; sternite 5 ( Fig. 16) a rectangular plate, slightly more sclerotised laterally; sternite 6 vague with a pair of small sclerotised sections (Fig. 16); synsternite 7+8 quite large, symmetrical, narrowing laterally, lateral slits enclosing hardly sclerotised areas (Fig. 17); both spiracles located in the lateral slits of the synsternite (Fig. 17); epandrium (Fig. 18) rounded, with ~ 11 pairs of setulae, ventrally bare, otherwise clothed in microtrichia; surstyli (Figs 19, 24) articulated, apically broadening, apex rounded and bulbous, with a long, apically broadening and rounded, apophysis; in posterior view (Fig. 24) a few small setulae on apical halves of surstylus and apophysis with the apical third of the surstylus and only the apex of the apophysis clothed in microtrichia, on inner side only microtrichia on the apices of surstylus and its apophysis with a few small setulae on apophysis and apical third of surstylus (Fig. 19); surstyli interconnected via thin, rod-like processus longi (Fig. 18); cerci rather broad, ratio of length/width 1.9, basally and apically tapering, apex rounded, clothed in microtrichia and a set of setulae, some of the apical setulae almost as long as the cerci (Fig. 18); hypandrial clasper (Fig. 20) straight and rod-like with relatively long setulae on distal half; phallapodeme solidly built and rather straight (Fig. 21), anterior arm rounded api- cally, posterior arm slightly longer than anterior arm (ratio posterior arm/anterior arm 1.05) and strongly bifurcated to accommodate the very broad phallus; phallus (Fig. 21) a rather short complex of lobes and sclerites, remarkably broad and heavily sclerotised, intromittent organ very short; ejaculatory apodeme straight, hardly broadening apically except for a small sickle-shape of apex (Fig. 22), ejaculatory sac rounded.
Distribution and habitat. The new species is only known from the Makay massif in Toliara province. The Makay is a mountain range of almost 4000 km 2 in southwestern Madagascar. The altitude varies from 200 m at the bottom of canyons to 1000 m for the plateaus. The Makay with its exceptional biodiversity (see Wendenbaum 2011) is considered to be one of the least studied areas in Madagascar. Its forests belong to the deciduous, seasonally dry, western forests of low altitude (Du Puy and Moat 1999). In the dry season, wet areas remain near the rivers. Many Diopsidae, including aggregations, were observed on vegetation in wet, shady places. The single specimen of this new species was collected in undergrowth along a riverbank at an altitude of 527 m. On the same location the following Diopsidae were collected: 5 ♀ and 5 ♂ Sphyracephala beccarii (Rondani), 5 ♀ and 2 ♂ Diopsis nigrosicus Séguy, and 1 ♂ M. parvapollina.
Etymology. This species is named M. makayensis Feijen, Feijen & Feijen, sp. nov., referring to the place of origin of the holotype.

Key to the species of Madagopsina
This key is a revised version of the Madagopsina section of the key in Feijen et al. (2018). It now also includes M. makayensis Feijen, Feijen & Feijen, sp. nov. for which only the male is known. The couplet separating the two species groups has been changed to accommodate the new species. In the 2018 key also an error occurred: in the couplet separating M. apollo and M. parvapollina, the character states for the apical seta should have been reversed.

Discussion
Geometric morphometrics analysis of wing shape Feijen et al. (2018) proved that principal component analysis (PCA) of wing morphometry was powerful enough to recover the Madagopsina and Gracilopsina taxa previously delimited by adult morphology. Their biplot of the first two PCA axes showed clear distinction between the two genera and seven species while these two axes explained 54.8% and 22.9% of variation. Only specimens from M. tschirnhausi and M. apographica overlapped slightly. The M. apographica species group and the M. apollo species group were also delimited. Feijen et al. (2018) stated that the limited intraspecific variation in the PCA plot could, to some extent, be explained by the fact that the wings in the, mostly, pinned flies are often not perfectly flat. So, the PCA leads to a cluster pattern that is in accordance with morphological characters. The same pattern was seen in the hierarchical clustering analysis of PCA scores. Only three of 68 specimens were assigned to the wrong species cluster using the 'complete' cluster method. Both analyses were now repeated with inclusion of the wing data for the single specimen of the new species. The biplot of the first two PCA axes (Fig. 29) shows the distinction between the two genera and eight species, while the two axes explained 54.7% and 22.7% of variation. M. makayensis Feijen, Feijen & Feijen, sp. nov. is placed squarely in the cluster for the M. apollo species group, while within this group it is more closely related to M. parvapollina than to M. apollo. This same pattern is seen in the hierarchical clustering analysis of PCA scores using the complete cluster method (Fig. 30). This method showed that within

Allometric aspects with regard to eye span
For all five Madagopsina species included in Feijen et al. (2018), graphs were presented for eye span plotted against body length for both sexes. The differences in allometric slopes for males and females indicated the rate of sexual dimorphism D for the species. Between the species the allometric slopes for males varied from 1.64-2.13 and for females from 0.85-1.21. Of special interest were the allometric lines for the two species then forming the M. apollo species group: M. apollo and M. parvapollina (Fig. 31). According to Feijen et al. (2018) the two species are externally very similar. They can in the first place be distinguished by the well-separated size ranges (see also Fig. 31).
Comparison of the allometric lines for the two species, showed that the female lines are collinear, but given the size difference well separated. In fact, the female lines for the M. apographica species group also do not differ much from those of the M. apollo group. The slopes of the male lines for M. apollo and M. parvapollina were almost similar with 1.87 and 1.89, respectively, but the intercepts are distinct. This leads to two parallel lines (Fig. 31). Now that M. makayensis Feijen, Feijen & Feijen, sp. nov. has become the third species in the M. apollo group, its place can be considered in the graph with the allometric lines for the other two species. Only one male specimen is available for M. makayensis Feijen, Feijen & Feijen, sp. nov., so a single data point is available in the graph (Fig. 31). The single male data point is in line with the allometric

Male genitalia
According to Feijen et al. (2018), preliminary results from divergence dating analysis suggest a minimum age estimate of around 10 million years for the divergence of M. freidbergi and M. tschirnhausi. However, reaching convergence in divergence dating analysis proved difficult and longer runs will later be required. Feijen et al (2018) considered the large differences in postabdominal structures in Madagopsina as additional support that its species diverged long ago. In M. makayensis Feijen, Feijen & Feijen, sp. nov., the male genitalia are also quite distinct from the other Madagopsina species. The major differences in surstyli for the six species are illustrated (Figs 23-28). In other Diopsidae genera and species groups, the differences in surstyli are often much smaller, as can, for instance, be seen in Eurydiopsis (Feijen 1999: figs 1-9), the sister genus of Madagopsina. The differences in hypandrial claspers are also large in Madagopsina, as can be noted by comparing these claspers in the new species (Fig. 20) with those of the other five species (Feijen et al. 2018: figs 178-182). For the closely related Syringogastridae, Marshall et al. (2009) referred to the hypandrial claspers as the "large, setulose ventral lobe" of the hypandrial arms. The short, broad, and heavily sclerotised phallus (Fig. 21) in M. makayensis Feijen, Feijen & Feijen, sp. nov. is not only unusual for Madagopsina, but for the whole Diopsidae family.

Morphological differences between the two species groups of Madagopsina
Due to the description of a new species in the M. apollo species group, the list of differences with the M. apographica group has to be somewhat revised. A major difference between the two groups, according to Feijen et al. (2018) concerns the anterior central hyaline wing spot. In the M. apollo group this spot was located in cells r1 and r2+3, while in the M. apographica group this spot only occurred in cell r2+3 and did not extend into cell r1. However, in M. makayensis Feijen, Feijen & Feijen, sp. nov. this spot also does not extend into cell r1, so this character can no longer be used to separate the two groups (compare Figs 9-11 with Figs 12-14). The slight difference in pruinescence of the tergites is now also removed from the list of differences. In the M. apollo group the tergites were glossy, while in the M. apographica group they were slightly pruinose. In M. makayensis Feijen, Feijen & Feijen, sp. nov. the tergites are also slightly pruinose, so this character is now also removed from the list of differences. In Table 1, the new list of differences is presented. Compared with Feijen et al. (2018), the range of some ratios is slightly adapted. In the wing pattern another major difference is now introduced: the width of the dark preapical wing band as compared with the wing length and also as compared with the central wing band. Within the M. apollo species group, M. parvapollina and M. makayensis Feijen, Feijen & Feijen, sp. nov. are more closely related to each other than to M. apollo. The latter species stands out by its much larger body size, much shorter inner vertical seta and apical seta, much shorter and bulbous apophysis of the surstylus, and its peculiar subanal plate (heart-shaped with bulbous lateral areas).

Conclusions
The division of Madagopsina into the M. apollo species group and the M. apographica species group is consolidated by the inclusion of the morphological data for M. makayensis Feijen, Feijen & Feijen, sp. nov. Within the M. apollo group, the new species is more closely related to M. parvapollina. The division into two species groups and the closer relationship of M. makayensis Feijen, Feijen & Feijen, sp. nov. to M. parvapollina is fully supported by the geometric morphometrics analysis of wing shape and the analysis of the allometric data with regard to eye span.