The Herniosina story continues in the Mediterranean: H.calabra sp. nov. from Calabria and H.erymantha Roháček, new female from the Peloponnese (Diptera, Sphaeroceridae)

Abstract A study of recently acquired material of Herniosina Roháček, 1983 (Diptera: Sphaeroceridae: Limosininae) in the Mediterranean subregion revealed a new species, H.calabrasp. nov. (Italy: Calabria: Serre Calabresi Mts) and the first females of H.erymantha Roháček, 2016 (Greece: southern Peloponnese: Taygetos Mts). Herniosinacalabrasp. nov. (both sexes) and the female of H.erymantha are described and illustrated in detail including structures of terminalia, their relationships are discussed and new information on their biology (habitat association) is given. An update of a key to all know species of Herniosina species is presented.


Introduction
The genus Herniosina Roháček, 1983 (Sphaeroceridae: Limosininae) has recently been reviewed by Roháček (2016), including all known species. In the latter study the genus was re-diagnosed without the Nearctic H. voluminosa Marshall, 1987 which has been removed so that the genus is monophyletic. Herniosina voluminosa was later transferred to a monobasic genus Voluminosa Roháček & Marshall, 2017 because it lacks the majority of the defining apomorphies of the Palaearctic Herniosina, see Roháček and Marshall (2017).
Herniosina was originally described during the re-classification of the previous genus Limosina Macquart, 1835 by Roháček (1982Roháček ( , 1983 for two European species, H. bequaerti (Villeneuve, 1917) and H. horrida (Roháček, 1978). A third species, H. pollex Roháček, 1993, was added by Roháček (1993) from Central Europe and two more species, H. erymantha Roháček, 2016 andH. hamata Roháček, 2016 have recently been described from the E. Mediterranean area (Greece and Cyprus) by Roháček (2016). In addition, there is a record of an unnamed species of Herniosina (based on two females) from Israel (Papp and Roháček 1988: 89, as Herniosina sp. cf. horrida). Discovery of new Herniosina species in the eastern Mediterranean (Roháček 2016) indicated that there could also be undescribed species in other parts of the area, and, therefore, our subsequent collecting trips to the Mediterranean have been focused on the acquisition of further material of this largely terricolous and/or subterranean genus. These collecting efforts resulted in two series of Herniosina specimens, one from southern Peloponnese (Taygetos Mts) and the other from Calabria (Serre Calabresi Mts). While the former proved to belong to H. erymantha and includes the first known females of the species, the other has been recognised to represent an unnamed species morphologically somewhat intermediate between H. erymantha and H. bequaerti.
The genus Herniosina has been newly diagnosed by Roháček (2016: 74-75). This diagnosis does not need a revision after the inclusion of the new species being described below. Herniosina species can be most easily recognised by a combination of (largely apomorphic) features in the male abdomen and terminalia (postabdomen strongly down-curved, S1+2 more or less bulging, S5 strongly reduced, cerci modified to peculiar projections, both distiphallus and phallophore projecting posteroventrally) and the (largely plesiomorphic) structures of the female postabdomen (narrow and telescopic, sclerites of 6 th and 7 th segments and also 8 th tergum well developed, no internal sclerites, cerci usually long and slender), however, with reduced S8 and S10.

Material
All the material examined (including specimens of Herniosina bequaerti used for illustrations but not listed below) is deposited in SMOC, Slezské zemské muzeum, Opava, Czech Republic.

Methods of preparation and study of postabdominal structures
Abdomens of some specimens were detached, cleared by boiling several minutes in 10% solution of potassium hydroxide (KOH) in water, then neutralised in 10% solution of acetic acid (CH 3 COOH) in water, washed in water and subsequently transferred to glycerine. Postabdominal structures were dissected and examined in a drop of glycerine under binocular microscopes (Reichert, Olympus). Detailed examinations of genital structures were performed with a compound microscope (Zeiss Jenaval). After examination, all dissected parts were put into small plastic tubes containing glycerine, sealed with hot forceps and pinned below the respective specimens. Specimens with abdomen removed and terminalia dissected are indicated in the list of material by the abbreviation genit. prep.

Drawing techniques and photography
Legs were drawn on squared paper using a Reichert binocular microscope with an ocular screen. Details of the male and female genitalia were drawn by means of Abbe's drawing apparatus on a compound microscope (Zeiss Jenaval) at larger magnification (130-500×). Wings were photographed on the compound microscope Olympus BX51 with an attached digital camera (Canon EOS 1200D). Whole adult (dry-mounted) specimens and wings were photographed by means of a digital camera Canon EOS 5D Mark III with a Nikon CFI Plan 10×/0.25NA 10.5 mm WD objective attached to a Canon EF 70-200 mm f/4L USM zoom lens. The specimen photographed by means of the latter equipment was repositioned upwards between each exposure using a Cognisys Stack-Shot Macro Rail and the final photograph was compiled from multiple layers (20-40) using Helicon Focus Pro 7.0.2. The final images were edited in Adobe Photoshop CS6.

Measurements
Six main characteristics of the new species were measured: body length (measured from anterior margin of head to end of cercus, thus excluding the antenna), index t 2 : mt 2 (= ratio of length of mid tibia : length of mid basitarsus), wing length (from wing base to wing tip), wing width (maximum width), C-index (Cs 2 : Cs 3 ) (= ratio of length of 2 nd costal sector : length of 3 rd costal sector) and index rm\dm-cu : dm-cu (= ratio of length of section between rm and dm-cu on discal cell : length of dm-cu). All type specimens and also all newly collected specimens of H. erymantha were measured.

Presentation of faunistic data
Label data of primary-type specimens are presented strictly verbatim including information on form and colour of all associated labels. Data from paratypes of the new species and also from formerly unpublished non-type specimens are standardised and presented in full. Phenological and other biological information obtained from the material examined and literature are given in the Biology paragraph.

Morphological terminology
Morphological terminology follows that used for Sphaeroceridae by Roháček (1998) in the Manual of Palaearctic Diptera including terms of the male hypopygium. The "hinge" hypothesis of the origin of the eremoneuran hypopygium, re-discovered and documented by Zatwarnicki (1996), has been accepted and, therefore, the following synonymous terms of the male genitalia (emanating from other hypotheses) need to be listed (terms used first): ejacapodeme = ejaculatory apodeme, epandrium = periandrium, medandrium = intraperiandrial sclerite, phallapodeme = aedeagal apodeme. with wing removed for photography and also preserved in glycerine in pinned plastic tube below the specimen (SMOC 06/001/2018-2 -06/001/2018-21). Etymology. The name of the new species is an adjective derived from Calabria (a region in southern Italy) where the type locality of the new species is situated.
Description. Male (Fig. 1). Total body length 2.06-2.46 mm; general colour blackish brown with mostly very sparse dark greyish brown microtomentum, hence body relatively shining. Head blackish brown to brown, lightest on gena. Frons blackish brown posteriorly to brown anteriorly, sparsely microtomentose and largely shining. Occiput blackish brown to black with sparse dark greyish brown microtomentum. Orbits, interfrontalia (very narrow, poorly delimited) and ocellar triangle also greyish brown to dark grey (orbits) microtomentose and duller than rest of frons; frontal triangle relatively narrow, glabrous and shining. Cephalic chaetotaxy: pvt absent, only minute adpressed postocellar setulae behind ocellar triangle; occe distinctly shorter than occi, the latter ~ 2/3 length of vte; vti longest among frontal setae, vte and oc slightly shorter than vti; two strongly exclinate and closely situated ors, posterior longer than anterior and both distinctly shorter than oc; 4 to (usually) 5 relatively short ifr, 1 or 2 middle pairs slightly longer than others; 4 very minute ads inside and below ors; g weak, hardly longer than anterior peristomal seta; vi long, ~ as long as vti. Frontal lunule short, wide, basally brown as anterior margin of frons, apically darkened. Face with cavities below antennae dark brown to black, shining despite sparse greyish microtomentum; medial carina distinct although slightly elevated. Gena high, brown in anterior half, blackish brown posteriorly, sparsely grey microtomentose. Eye relatively small; its longest diameter ~ 1.9 × as long as smallest genal height. Antenna relatively long, black or 3 rd segment blackish brown; 3 rd segment distinctly tapered apically both in dorsal and lateral view, with cilia on apex as long as those longest on arista. Arista long, ~ 3.8 × as long as antenna, in basal 1/4 short ciliate, otherwise moderately long ciliate.
Thorax dark brown to black, mesonotum relatively shining because of sparse microtomentum, pleuron more densely microtomentose and duller (Fig. 1). Some su- tures between pleural sclerites pale brown. Scutellum relatively large and long, rounded triangular, with dorsal surface flat and finely microsculptured, duller than mesonotum. Thoracic chaetotaxy: 2 hu but internal reduced to microseta; 2 postsutural dc, anterior short and weak (only 2 × longer than dc microsetae), posterior strong, ~ as long as or slightly shorter than basal sc; 8-10 rows of ac microsetae on suture; medial prescutellar ac pair somewhat prolonged and thickened but shorter than anterior dc; 2 strong sc, basal slightly longer than scutellum, apical (longest thoracic seta) ~ 1.5 × as long as basal; only 1 stpl because anterior stpl reduced to hardly discernible microseta.
Legs dark brown, coxae, trochanters, knees and tarsi brown to pale brown. f 1 with sparse and relatively short setae in posterodorsal and posteroventral rows. f 2 with a row of 4-6 curved but relatively short ventral setae in basal third (Fig. 5) in addition to the usual fine basal seta; t 2 ventrally with a long row of small dense spines terminated by a strongly reduced va seta (markedly shorter than anteroapical seta), see Fig. 5; dorsal chaetotaxy of t 2 as in congeners including relatively variable-in-length posterodorsal seta in apical fourth (Fig. 4). t 2 : mt 2 = 1.91-2.02.
Wing ( Fig. 2) with pale brownish membrane and pale brown to blackish brown veins. C hardly produced beyond apex of R 4+5 . R 2+3 slightly sinuate to straight but apically distinctly upcurved to C; R 4+5 sinuate but with apical half almost straight. Discal cell (dm) variable, relatively short to medium long, distally more or less tapered, usually with small process of M beyond dm-cu (venal fold of M continuing this process usually well visible); posterior outer corner of dm cell obtuse-angled, often with small to minute process of CuA 1 beyond dm-cu, rarely rounded (1 specimen). A 1 slightly sinuate; anal lobe well developed; alula narrow but not acute. Wing measurements: length 1.88-2.32 mm, width 0.77-0.97 mm, C-index = 0.87-1.17, rm\dm-cu : dmcu = 2.87-3.67. Haltere with dirty yellow stem and dark brown knob.
Postabdomen (Figs 17-19) telescopically retractable, basally (6 th segment) markedly narrower than preabdomen at 5 th segment. 6 th segment (both T6 and S6) distinctly wider than 7 th segment in contrast to those of H. bequaerti. T6 wide and short, transversely trapezoidal, with pale-pigmented anterior and (wider) posterior marginal stripe (Fig. 17), setose at lateral and posterior margins, with longest setae in posterior corners; T7 distinctly narrower than T6 and reaching farther onto lateral side (Fig. 19), with small unpigmented anteromedial area and setosity restricted to posterior margin (Fig. 17). T8 as long as T7 but dorsomedially narrowly depigmented and appearing divided into two dark sclerites (Fig. 17), in contrast to T8 of both H. bequaerti and H. erymantha. T10 transversely subtriangular (Fig. 17), shorter than those of H. bequaerti and H. erymantha), pigmented (darkest anterolaterally) except for posterior corner, with a pair of long setae, some fine setulae and micropubescent on almost entire surface. S6 somewhat wider, shorter (more transverse), slightly paler and more setulose than S7 (Fig. 18). S7 dark-pigmented except for posterior marginal stripe and with 4 longer and several short setae at posterior margin. S8 (Figs 18, 19) reduced, short but wider than those of H. bequaerti and H. erymantha, strikingly convex at anterior margin where densely micropubescent (cf. Fig. 19), otherwise with only 6-8 short setae. S10 reduced to distinctive transverse (in ventral view sinuous) sclerite, being medially depigmented (Fig. 18) but laterally blackish brown and posterodorsally rectangularly incised (Fig. 19), which is also visible in dorsal view (Fig. 17). S10 densely micropubescent and with a few setae including 1 long pair. Spermathecae 2+1 (Fig. 20) blackish brown, pyriform with conical bases, most resembling those of H. erymantha, sharing with the latter distally ringed conical bases, dark thickened apex and terminal parts of ducts of paired spermathecae connected rather far from their bodies; however, spermathecae of H. calabra are more robust, with wider basal conical parts. Cerci (Figs 17-19) more robust than those of H. bequaerti but much longer and narrower than those of H. erymantha, each bearing 1 dorsal preapical and 1 apical setae, both very long and sinuate, apart from other shorter setosity and dense micropubescence.
Remarks. Herniosina calabra sp. nov. seems to be morphologically intermediate between H. bequaerti and H. erymantha. Although seemingly more similar to H. erymantha (smaller body size, shorter male T5 and S8, male S5 with deeply forked medial process, anterior process of male cercus long and robust, gonostylus ventrally rounded, not emarginate, spermathecae with conical basal part distally ringed) it is probably most closely related to H. bequaerti. Sister-species relationship of H. calabra and H. bequaerti seems to be particularly demonstrated by the following putative synapomorphies: very similar construction of the male aedeagal complex, including the short distiphallus (with both lateral lobes and unpaired ventral process short) and surprisingly similarly formed, short and robust postgonite. In the female postabdomen there is also a shared synapomorphy: the modified (posterodorsally more or less incised) lateral part of S10 (cf. Fig. 19 and Fig. 42).
The new species can be easily separated from all known congeners only by postabdominal characters. The most species-specific are as follows: the long, slender and deeply forked medial process of male S5 (Fig. 7); the male S8 with digitiform lobes on both sides of lateral slit (Fig. 3); the gonostylus with a posteromedial tooth (Fig.  10), the male cercus with anterior (more lateral) lobe with apex bent outwards (Fig.  9); the postgonite short and with robust apex (Fig. 13); the lateral part of female S10 dark-pigmented and with posterodorsal rectangular incision (Fig. 19). Moreover, the combination of female T8 medially narrowly depigmented with short T10 and relatively long slender cerci (see Fig. 17) is also very characteristic.
Biology. The entire type series of H. calabra sp. nov. (21 specimens) was collected (aspirated by a pooter) in May under Juncus tufts (Fig. 22) growing under alder trees surrounding a small creek in a montane meadow (Fig. 21). The sphaerocerid community co-occurring with H. calabra in and under these tufts of rush (based on collected specimens) proved to be relatively rich and contained the following 15 species:  Roháček, 1983 1♂ andTerrilimosina schmitzi (Duda, 1918) 1♀. This assemblage included largely saprophagous terricolous species (such as H. calabra, G. flaviceps, Limosina silvatica, Pteremis fenestralis, Pullimosina species, Puncticorpus cribratum, T. schmitzi) but also a few microcavernicolous species (Spelobia talparum, S. sp. cf. talis) and some ubiquitous, predominantly coprophagous, species (Lotophila atra, Sphaerocera curvipes, Spelobia clunipes). The presence of the latter two groups indicates that there could also be some droppings of small mammals in the detritus. This is for the first time that a species of Herniosina has been found under tufts of a graminoid plant. However, rotting leaves of alder were also present under tufts of Juncus sp. ex- amined (see Fig. 22), which indicate more resemblance to a leaf-litter association as known in most other Herniosina species (cf. Roháček 2016).
Distribution. Hitherto only known from S. Italy (Calabria).
Supplementary description. Male (Fig. 23). Total body length 1.79-2.46 mm. Head. Cephalic chaetotaxy: 3 or 4 relatively short ifr, subequal in length or the middle pair longer. Gena high, usually reddish-brown only anteriorly, sometimes on most of genal surface. Third antennal segment with ciliation on apex as long as longest cilia on arista.
Thorax. Scutellum relatively large and long (1.5 ~ as wide as long), rounded triangular, with dense fine microsculpture on flat dorsal surface. Thoracic chaetotaxy: 1 or 2 stpl, posterior long, anterior reduced to microseta or absent.
Legs. f 2 with a long row of 6-8 curved but relatively short ventral setae in basal half to two-thirds. t 2 : mt 2 = 1.83-1.90.
Wing. Discal cell (dm) variable, relatively short to medium long, distally usually less tapered than in most relatives, with small process of M beyond dm-cu being continued by a venal fold; posterior outer corner of dm obtuse-angled to rounded, sometimes with small remnant of CuA 1 . Wing measurements: length 1.87-2.38 mm, width 0.77-1.01 mm, C-index = 0.88-1.09, rm\dm-cu : dm-cu = 2.62-3.15.
Genitalia. Epandrium besides a group of longer and stronger setae laterally and lateroventrally usually also with 1 longer dorsolateral seta which can sometimes be reduced (as is in the holotype, see Roháček 2016: figs 20, 21). Gonostylus (Fig. 41) with posterodorsal corner broadly rounded, never tooth-like and projecting.
Postabdomen  telescopically retractable but broader than in H. bequaerti or H. calabra, particularly as regards 7 th and 8 th segments when compared with width of 5 th abdominal segment. T6 wide and short, transversely oblong, with pale-pigmented posterior marginal stripe (Fig. 27), sparsely setose at posterior and lateral margin, with 1 long seta in each posterior corner; T7 only slightly narrower than T6 (Fig. 27) but reaching farther onto lateral side (Fig. 29), sparsely setose only at posterior margin and with very narrowly unpigmented posterior margin. T8 shorter and narrower than T7, all darkpigmented or only narrowly paler at posterior margin medially (Fig. 27). T10 shortly pentagonal, rounded laterally, less transverse than that of H. calabra but shorter than that of H. bequaerti, pale-pigmented only anteriorly and laterally, and dorsally with a pair of long setae, a few fine setulae and entirely covered by micropubescence (Fig. 27). S6 wider, more transverse and more densely setulose than S7, dark-pigmented except for posterior margin (Fig. 28), with 4 or 6 long posterior setae. S7 also dark but with narrowly unpigmented anterior margin (Fig. 28) and with 4 long (those in medial pair close to each oth-  er) setae in addition to sparse short setae in posterior half. S8 (Figs 28,29) small, narrower than that of H. calabra, having posterior half tapered, with several fine setae (4 longer) and distinctive micropubescence, particularly anteromedially. S10 reduced to short, Vshaped, micropubescent and setose sclerite being medially depigmented to interrupted (Fig. 28), with lateral pigmented parts simple (Fig. 29) in contrast to those of H. calabra. Spermathecae 2+1 (Fig. 30) blackish brown, elongate pyriform, most resembling those of H. calabra but with basal conical parts narrower. Cerci (Figs 27-29) markedly different from those of both H. calabra and H. bequaerti, unusually short and robust (more so than in H. hamata Roháček, 2016), apically conical and dorsoventrally somewhat flattened, each with 1 dorsal preapical and 1 apical seta long sinuate and 1 ventral preapical seta curved (apart form a number of shorter setae), and with dense micropubescence.
Remarks. Herniosina erymantha sp. nov. has only been known from the male holotype (Roháček 2016). A series of specimens recorded here enabled the description of the male to be supplemented and to add the first description of the female. As mentioned above (see Remarks under H. calabra), H. erymantha seems to be most closely allied to the sister-pair H. bequaerti -H. calabra. This relationship can now also be confirmed by the female postabdominal characters, including the similar formation of female S8 and, particularly, by the medially depigmented (to almost interrupted) S10 (cf. Fig. 28).
On the other hand, female H. erymantha can be easily distinguished from females of both its relatives (and also from all other congeners) by the unusually robust cerci (Fig. 27) and the detailed shape of S8 and S10 (Fig. 28).
Biology. Almost all newly obtained specimens of H. erymantha were swept from above decaying leaf-litter and sparse vegetation under Platanus trees in valleys of montane brooks in the Taygetos Mts (Figs 31, 32), usually mostly in humid places (shores of brooks, springs). Because the holotype was sifted from dead leaves of Platanus in a similar montane habitat in the Erimanthos Mts (see Roháček 2016) it is very probable that its larvae develop in this microhabitat. Adults are now known to occur in May (Roháček 2016) and October (present data).
Distribution. Hitherto only known from Greece: Peloponnese.

Discussion
The monophyly of Herniosina was demonstrated by Roháček (2016) and its affiliation within the Limosina group of genera (sensu Roháček 1982) was confirmed by Roháček and Marshall (2017 ence of some putative synapomorphies of Volumosina Roháček & Marshall, 2017 and Apteromyia (cf. Roháček and Marshall 2017: 459), the sister group relationship of Herniosina and Apteromyia remains as the most probable (Roháček 2016). Particularly, the peculiar modifications of the aedeagal complex (distiphallus with unpaired ventromedial lobe projecting posteriorly; phallophore anteriorly slender and elongate, projecting, movably attached to dorsal side of distiphallus) and the male cerci (formed as robust compact processes below anal fissure) are possible unique synapomorphies of these two genera. The addition of Herniosina calabra sp. nov. did not affect the concept of the genus which remains a compact group of similar species, differing mainly by the structures of the male and female terminalia. This new species seems to be most closely allied to H. bequaerti and H. erymantha, and, consequently the hypothesis of the relationships of species within Herniosina has to be changed as follows. Herniosina hamata is considered a sister-taxon to the five other congeners which belong to a monophyletic group supported by the following putative synapomorphies: male preabdominal sclerites with setosity reduced; male cercus modified to 2 (lateral and medial) processes; gonostylus with dorsal internal projection; spermathecae pyriform, with distinct conical basal part. Herniosina pollex, with the bulge of the male S1+2 small (a plesiomorphy shared with H. hamata) is considered a sister-group to a clade with H. horrida, H. erymantha, H. calabra and H. bequaerti being supported by 2 synapomorphies (male S1+2 strongly bulging; gonostylus with slender dorsal internal projection). Herniosina horrida (having male S5 with a pair of small posteromedial projections = a plesiomorphy shared with H. hamata and H. pollex) seems to branch off the remaining triplet formed by H. erymantha, H. calabra, and H. bequaerti. The close alliance of these three species