Herniosina Roháček: revised concept, two new species, new key and atlas of male and female terminalia (Diptera, Sphaeroceridae)

Abstract The taxonomic concept of Herniosina Roháček, 1983 (Diptera: Sphaeroceridae) is revised on the basis of five W. Palaearctic species, thus excluding the E. Nearctic Herniosina voluminosa Marshall, 1987 whose inclusion caused the paraphyly of the genus. Two new species, Herniosina erymantha sp. n. (male only, Greece: Peloponnese) and Herniosina hamata sp. n. (both sexes, Cyprus), are described and illustrated, and the other three species, Herniosina bequaerti (Villeneuve, 1917), Herniosina horrida (Roháček, 1978) and Herniosina pollex Roháček, 1993, are diagnosed with an atlas of their male and female terminalia. The relationships of the redefined genus and of all its species are discussed, and their biology and distribution are reviewed. A new illustrated key to Herniosina species is given.


Introduction
The genus Herniosina was established by Roháček (1982Roháček ( , 1983 during the re-classification of the giant assemblage previously included in the genus Limosina Macquart, 1835 to comprise two European species of the subfamily Limosininae, viz. Herniosina bequaerti (Villeneuve, 1917) and H. horrida (Roháček, 1978). The genus was characterized by the peculiar down-curved male abdomen with protruding bulge on abdominal synsternum S1+2 (cf. Fig. 1) and distinctive male genitalia. Discussing its relationships Roháček (1982) suggested it obviously belongs together with the monotypic genera Limosina (redefined), Gigalimosina Roháček, 1983 andApteromyia Vimmer, 1929 to the Limosina genera-group. When describing Apteromyia newtoni (a second species of the genus) Marshall and Roháček (1982) hypothesized Apteromyia as the closest relative of Herniosina. Marshall (1987) described another unusual species of Limosininae and placed it tentatively in Herniosina on the basis of (seemingly) similarly protruding male abdominal S1+2 despite his explicit recognition that it was markedly different in other characters of the male genitalia, i. e. lacking all other synapomorphies of the genus as originally delimited. The inclusion of H. voluminosa Marshall, 1987 made Herniosina a heterogeneous and apparently non-monophyletic group, as was stated already by Roháček (1993) when describing the third European species, H. pollex Roháček, 1993. Apart from a record of  (Villeneuve), male laterally (Czech Republic: Bohemia). Body length cca 2.6 mm. S1+2 and T5 arrowed. Photograph by J. Roháček. two unidentified females from Israel (Papp and Roháček 1988: 89, as Herniosina sp. cf. horrida) no further species of Herniosina had been known until Roháček (2004) reported about an unnamed species found in Cyprus. However, the latter species has remained undescribed up to the present. Only the recent (2015) discovery of an additional new species in the Peloponnese peninsula initiated the present study which is not only aimed at the descriptions of these two new species but also at revision of taxonomic limits of the genus, its re-definition on the basis of the study of all known species, hypotheses of their relationships as well as at an updated synopsis of their biology and distribution.

Material
The material examined is deposited in institutional and private collections as follows: of genital structures were performed with a compound microscope (JENAVAL). After examination, all dissected parts were put into small plastic tubes containing glycerine, sealed with hot forceps and pinned below the respective specimens.

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 (JENAVAL) at larger magnification (130-500×). Wings were photographed on the same microscope with an attached digital camera (Nikon COOLPIX 4500). Whole specimens were photographed by means of digital camera Canon EOS 5D Mark III with macro lens Canon MP-E 65 mm 1-5× and ring macro flash Canon MR-14EX.

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 2nd costal sector : length of 3rd 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 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 standardized and presented in full. For data of paratypes or paralectotypes of other species and formerly published records original publications are cited. Phenological and other biological information obtained from the material examined and literature are given in the Biology paragraph; data on distributions are presented as summarized by  and Marshall et al. (2011).
Discussion. The genus Herniosina can be identified by the key to European (Roháček 1983) and/or Palaearctic Limosininae (Roháček 1998). It seems to be best recognized by combination of apomorphic characters in the male abdomen and terminalia (postabdomen strongly down-curved, S1+2 bulging, S5 strongly reduced, cerci modified to peculiar projections, both distiphallus and phallophore projecting posteroventrally) and the plesiomorphic formation of the female postabdomen (relatively narrow and telescopic, no internal sclerites) having reduced S8. Although the diagnosis of the genus has to be somewhat modified (see above) with respect to the inclusion of the two new species described below, the genus remains to be a very compact monophyletic group if only Palaearctic species (reviewed here) are included.
However, Marshall (1987) described a very peculiar Limosinine species from USA (New Hampshire) and placed it as a tentative member of Herniosina on the basis of its bulging male S1+2 although this Herniosina voluminosa Marshall, 1987 differs very markedly from all other Herniosina species in many characters, lacking all other synapomorphies of the genus as originally delimited, including the modified male preabdominal sterna, reduced male S5, enlarged T5, projecting male cerci, form of phallophore, general shape of spermathecae etc. I have examined male and female paratypes of H. voluminosa kindly donated by S. A. Marshall to SMOC (USA: New Hampshire: Coos Co., 3 mi. NE East Inlet Dam, Norton Pool, flight interception trap, 12.-24.vi.1986trap, 12.-24.vi. , 1♂, 25.vi.-9.vii.1986 and found that the male S1+2 of H. voluminosa is differently (posteromedially) protruding and somewhat bilobed (see also Marshall 1987: Fig. 1) suggesting that this modification of S1+2 evolved independently and hence cannot be a synapomorphy of this Nearctic species and Herniosina s. str. species. Also its extremely enlarged aedeagal complex (with enormous distiphallus, phallophore and postgonite being several times larger than epandrium) and quite differently formed female terminalia (with S8 large and transverse, internal spectacles-shaped sclerite developed, spermathecae elongate and transversely wrinkled and T10 fused with base of cerci) clearly demonstrate that these taxa cannot be congeneric. Inasmuch as it seems that both known Apteromyia species (cf. Marshall and Roháček 1982) are apparently more closely allied to the Palaearctic species of Herniosina than is H. voluminosa (see below) the latter species is excluded from Herniosina here to render the genus monophyletic. It is therefore suggested to establish a new genus for the removed H. voluminosa in the near future.
When describing the genus, Roháček (1982Roháček ( : 221, 1983 placed Herniosina in the Limosina genera-group and discussed its affinity as being either a sister-group of the genus Apteromyia Vimmer, 1929 or sister-group of a clade comprising besides the latter genus also two other members of the Limosina genera-group, viz. Limosina Macquart, 1835 and Gigalimosina Roháček, 1983. Subsequently, Marshall & Roháček (1982 considered the former hypothesis to be more probable pointing out similarly modified (projecting) male cercus and setosity of epandrium in the Nearctic species Apteromyia newtoni Marshall & Roháček, 1982 and Herniosina species. Based on the present study the relationships of Herniosina and Apteromyia is supported by four supposedly synapomorphic characters (all in the male genitalia): epandrium with a series of robust ventral lateral setae; male cerci modified to compact processes below anal fissure, distiphallus with unpaired ventromedial lobe projecting posteriorly; phallophore anteriorly slender and elongately projecting, movably attached to dorsal side of distiphallus.
Because of being adapted to cold and the complete darkness in aphotic parts of caves and having the ability to develop under these conditions, Papp and Plachter (1976) classified H. bequaerti as a troglophilous species. They found larvae consuming various  (Villeneuve), male (Czech Republic: Bohemia). 2 Abdomen, laterally 3 Gonostylus, laterally 4 Genitalia, laterally 5 External genitalia, caudally 6 Aedeagal complex (phallapodeme partly omitted), laterally. Scales: 0.5 mm (2), 0.1 mm (3, 6), 0.2 mm (4, 5). For abbreviations see pp. 73-74. Adapted from Roháček (1978). (Villeneuve), female, male (Czech Republic: Bohemia). 7 Female postabdomen, dorsally 8 Ditto, ventrally 9 Ditto, laterally 10 Spermathecae 11 Male S5, ventrally. Scales: 0.2 mm (7-9), 0.1 mm (10, 11). For abbreviations see pp. 73-74. Adapted from Roháček (1978Roháček ( , 1983. decaying media, viz. dead animals, rotten vegetation including wood and/or mycelia of fungi and determined the length of its life-history (from egg to imago) within caves as 70-90 days and the life as an adult being 38 days on the average. The sometimes abundant occurrence of the species in cellars (see Roháček 1978 and material examined), caused by convenient conditions and a rich supply of larval food, can be considered a special case of synanthropy (Roháček 1983). In burrows of mammals the larvae develop in their droppings and other nest debris (Rotheray 1991) but obviously much more rapidly due to higher temperature. Adults can occur throughout the whole year. Puparia of H. bequaerti were described by Goddard (1938) and Skidmore (1993) based on specimens found in mouse-runs and on subfossil specimens excavated in an archaeological site in Iceland respectively.
Etymology. The name of the new species is an adjective derived from the Erimanthos (= Lat. Erymanthos) Mts inasmuch as its type locality is situated in the western part of this montane range.
Description. Male. Total body length 1.79 mm; general colour blackish brown with relatively sparse dark greyish brown microtomentum. Head blackish brown to brown. Frons largely blackish brown, brownish only at anterior margin, rather sparsely microtomentose. Occiput blackish brown and dark greyish brown microtomentose. Orbits, interfrontalia (poorly delimited) and ocellar triangle also greyish brown (not densely) microtomentose and duller than rest of frons; frontal triangle comparatively wide and shining. Cephalic chaetotaxy: pvt absent, only minute adpressed postocellar setulae behind ocellar triangle; occe and occi subequal and less than half length of vte; vti longest among frontal setae, vte and oc slightly shorter than vti; 2 strongly exclinate ors, both distinctly shorter than oc; only 3 relatively short ifr, the middle pair longest; 3-4 very minute ads inside and below ors; g weak, not longer than anterior peristomal setula; vi about as long as vti. Frontal lunule short, wide, similarly brown as anterior margin of frons. Face with cavities below antennae dark brown, rather shining; me-dial carina slightly elevated but distinct. Gena high, reddish only anteriorly, otherwise blackish brown, sparsely greyish microtomentose. Eye relatively small; its longest diameter about 2.2 times as long as smallest genal height. Antenna blackish brown, relatively long; its 3rd segment distinctly tapered apically both in dorsal and lateral view. Arista long, about 3.8 times as long as antenna, relatively long ciliate.
Thorax dark brown to blackish, mesonotum subshining due sparser microtomentum, pleuron lighter and dull. Suturae between pleural sclerites paler brown. Scutellum relatively large and long, rounded triangular. Thoracic chaetotaxy: 2 hu but internal reduced to microseta; 2 postsutural dc, anterior short and weak (only twice longer than ac microsetae), posterior strong but slightly shorter than basal sc; 8-10 rows of ac microsetae on suture; medial prescutellar ac pair somewhat prolonged; 2 strong sc, basal slightly longer than scutellum, apical about 1.6 times as long as basal; 2 stpl but anterior reduced to minute hair-like setula.
Legs dark brown, coxae, trochanters, knees and tarsi pale brown to ochreous. f 1 with relatively sparse setae in posterodorsal and posteroventral rows. f 2 with a row of curved but relatively short ventral setae in basal half (Fig. 18); t 2 ventrally with a long row of small dense spines and reduced va seta (shorter than anteroapical seta), see Fig.  18; dorsal chaetotaxy of t 2 as in congeners but posterodorsal seta in apical fourth somewhat shorter (Fig. 16). t 2 : mt 2 = 1.84.
Wing (Fig. 14) with pale brownish membrane and ochreous to dark brown veins. C hardly produced beyond apex of R 4+5 . R 2+3 slightly sinuate but apically distinctly upcurved to C; R 4+5 sinuate and its apical half almost straight. Discal cell (dm) relatively short and distally less tapered than in most relatives, with small process of M beyond dm-cu; posterior outer corner of dm not rounded but obtuse-angled. A 1 sinuate; anal lobe well developed; alula narrow but not acute. Wing measurements: length 1.87 mm, width 0.77 mm, C-index = 0.95, rm\dm-cu : dm-cu = 2.62. Haltere with ochreous stem and dark brown knob.
Genitalia. Epandrium (Figs 20, 21) of medium length but comparatively wide and more angular dorsolaterally (see Fig. 21) than those of relatives, and with a group of Herniosina erymantha sp. n., male holotype and its habitat (Greece: Peloponnese). 12 Valley above Alepochori in Erimanthos Mts, habitat in the type locality 13 Microhabitat (arrow) from where the holotype was sifted 14 Right wing (length 1.87 mm). Photographs by J. Roháček. longer and stronger setae laterally and lateroventrally (posterior seta longest and most robust) but dorsolaterally without longer seta. Anal fissure slightly wider than high (Fig. 21), subcircular. Cerci fused with epandrium, posteroventrally projecting in 2 processes, one (more anterior) robust, long (about as long as gonostylus) and distally somewhat thickened and bearing 1 long seta in addition to a number of setulae, the other (more medial) small, elongately conical and simple (Figs 20,21 small (low), reduced but connected by long internal arms with gonostyli (Fig. 21), posteromedially fused with cerci. Hypandrium with very long and slender anteromedial rod-like apodeme (Fig. 20). Gonostylus (Figs 20-23) most resembling that of H. bequaerti but ventrally simple (not emarginate in lateral view), with longer anteroventral seta, and with long, very slender, curved and apically lancet-shape dorsal internal process (cf. Fig. 22). Aedeagal complex (Figs 24-26) with long phallapodeme (as in H. bequaerti) but having distinctly lower dorsal keel. Aedeagus also somewhat similar to that of H. bequaerti but with funnel-shaped apex of distiphallus more robust, its lateral lobes and unpaired ventral process markedly longer and its postgonite more slender, with curved, slender and terminally blunt apex. Phallophore closely resembling that that of H. bequaerti, anteriorly rod-like but dorsoventrally flattened, posteriorly projecting ventrally and hence epiphallus-like. Ejaculatory apodeme not observed.
Female unknown. Discussion. Herniosina erymantha sp. n. is only known from the male holotype and, consequently, for the evaluation of its relationships the female characters cannot be used. However, based on the male terminalia the species is a distinctive member of Herniosina unmistakeably recognizable from any other known congener. It differs from all relatives by the flattened, distinctively forked medial process of S5 (Fig. 19), broad epandrium with strikingly rectangular dorsolateral corners without dorsolateral long seta and with broad (wider than high) anal fissure (Fig. 21), relatively simple sub-oblong gonostylus ( Fig. 23) with very slender and long internal process (Fig. 22), distinctive cercus with long and robust, distally somewhat dilated but laterally flattened lateral process combined with short medial ones (Figs 20, 21) and characteristic postgonite (Fig. 24).
Judging from the construction of the male abdomen and male genitalia, H. erymantha seems to be related to H. bequaerti and H. horrida sharing with them the following synapomorphies: strongly bulging S1+2; very slender and dorsally situated internal process of gonostylus. Its closest relative obviously is H. bequaerti having similarly (albeit much more) prolonged and basally fused medial processes of S5, a small medial apically pointed process of cercus and more robust funnel-shaped apex of distiphallus.
Biology. The holotype of H. erymantha sp. n. was sifted towards the end of May from dead leaves under Platanus trees (see Fig. 13, arrow) in a valley of a montane brook in the western ridge of the Erimanthos Mts (Fig. 12). The microhabitat (layers of decaying leaves of broad-leaved trees) is similar to that known for H. horrida and H. pollex in Central Europe (see under these species).
Distribution. Hitherto only known from Greece: NW Peloponnese.
Biology. The species is associated with decayed herbaceous vegetation, most of known specimens having been collected from decayed leaves, hay, grass (Roháček 1978(Roháček , 1983 in forests or their margins, in cold montane valleys also in open, unforested habitats; they can be also captured by soil traps in these habitats (Roháček 1980). Grundmann (1991) collected a series by Barber traps. Only occasionally it can be found in forests on excrement, rotting fungi (Roháček 1993), in runs of voles (Roháček 2009) and recently was also found in entrances of caves (Roháček 2011). Despite the latter record, the statement by Roháček (1983) that it has never been collected in caves together with H. bequaerti remains to be correct. However, H. horrida could possibly co-occur with H. pollex in mouths of caves or in ravines but hitherto I cannot confirm this presupposition by records. Adults were recorded in March to August.
Etymology. The new species is named by the adjective "hamata" to reflect its hook-like posteromedial (internal) process of the gonostylus (hamatus = Lat. provided with hooks).
Description. Male (Fig. 47). Total body length 1.94-2.38 mm; general colour blackish brown with relatively sparse dark greyish brown microtomentum, subshining (thorax) to shining (abdomen). Head blackish brown to brown, much higher than long (Fig. 47). Frons brown anteriorly, dark brown to blackish posteriorly, sparsely microtomentose. Occiput blackish brown with microtomentum denser. Orbits, interfrontalia and ocellar triangle sparsely greyish brown microtomentose; frontal triangle poorly delimited, relatively wide, anteriorly acute and more shining than rest of frons. Cephalic chaetotaxy: pvt absent, only minute divergent postocellar setulae behind ocellar triangle; occe and occi subequal (or occi slightly longer) and about half length of vti; vti robust and longest of frontal bristles; vte and oc slightly to distinctly shorter than vti; 2 ors, posterior longer than anterior but distinctly shorter than oc; 3-4 relatively short ifr, all subequal or foremost shorter, if 4 ifr present, the foremost reduced to small setula; 3-6 very minute ads inside and below ors; g weak, hardly longer than anterior peristomal setula; vi as long as or longer than vte but thinner. Frontal lunule short and wide, brown and greyish brown microtomentose. Facial cavities below antennae dark brown, relatively shining; medial carina poorly developed but distinct. Gena brown (lightest on virbrissal angle), posteriorly dark brown, greyish brown microtomentose. Eye relatively small; its longest diameter about 2.2 times as long as smallest genal height. Antenna dark brown, relatively long; its 3rd segment distinctly tapered apically, both in lateral and dorsal view. Arista long, about 4 times as long as antenna, relatively long and densely ciliate.
Thorax blackish brown and dark greyish brown microtomentose, mesonotum subshining, pleuron with denser microtomentum and dull. Suturae between pleural sclerites pale brown. Scutellum large, relatively long and flat, rounded trapezoidal. Thoracic chaetotaxy: 2 hu, internal reduced to microseta; 2 postsutural dc, anterior short (only 2-3 times as long as dc microsetae), posterior as long as or slightly longer than basal sc; 10 dense rows of ac microsetae on suture; medial prescutellar ac pair prolonged, about as long as anterior dc but finer; 2 long sc, basal about as long as or slightly longer than scutellum, apical about 1.7 times as long as basal; 1-2 stpl, anterior (if present) reduced to very minute pale setula.
Legs brown, coxae, trochanters, knees and tarsi paler brown to ochreous. f 2 with a doubled row of curved but relatively short ventral setae in basal half (Fig. 54); t 2 ventrally with a long row of small dense spines, very reduced va seta and 1 distinct anteroapical seta (Fig. 54); dorsal chaetotaxy of t 2 as in Fig. 57, with posterodorsal seta in apical fourth long. t 2 : mt 2 = 1.85-1.89.
Wing (Fig. 48) with pale brown membrane and yellowish brown to dark brown veins. C ending at or very slightly produced beyond apex of R 4+5 . R 2+3 very slightly sinuate and also apically slightly upcurved to C; R 4+5 distinctly sinuate but its apical half nearly straight. Discal cell (dm) rather long, distally tapered, with small process of M beyond dm-cu; posterior outer corner of dm cell varies from angular with a remnant of CuA 1 to rounded and lacking the latter. A 1 slightly sinuate; anal lobe large, well developed; alula narrow, apically rounded. Wing measurements: length 2.06-2.24 mm, width 0.85-0-91 mm, C-index = 0.97-1.03, rm\dm-cu : dm-cu = 3.64-4.50. Haltere with yellowish brown to ochreous yellow stem and dark brown knob.
Abdomen blackish brown dorsally, brown ventrally. Preabdominal terga large, glabrous and shining because of reduced and sparse greyish microtomentum (restricted to T1+2 and bases of T3-T5), sparsely setose but with more setae on disc than those of relatives. T5 enlarged but not so prolonged as in H. bequaerti. Preabdominal sterna S1+2-S4 (Fig. 51) also shining but distinctly more setose than in all congeners. S1+2 with ventral bulge reduced (Fig. 47, arrow) but somewhat protruding in the middle; S3 and S4 (Fig. 51) anteromedially emarginate and with characteristic lateral dark pigmentation. S5 (Fig. 65) very shortened, reduced to a transverse strip-like and largely weakly sclerotized and pale-pigmented sclerite with only partly darker and setose lateral parts (with 5-6 setae on each side) and posteromedially provided with a pair of small digitiform processes arising from a common base. S6+7 and S8 asymmetrical on left side of down-curved postabdomen, most similar to those of H. pollex, thus the latter relatively short.
Genitalia. Epandrium (Figs 49, 50) of medium length but comparatively broad (wider than high), with a series of longer and stronger setae mainly lateroventrally (posterior longest); also laterodorsally with 1 longer seta. Anal fissure relatively small, suboval, slightly wider than in H. pollex. Cerci fused with epandrium, each projecting ventrally in single (medial process absent) long, digitiform, terminally somewhat dilated process with blunt apex, basally carrying 1 long seta (Figs 49, 50). Medandrium fused with cerci medially and connected by long internal arms with gonostyli (Fig. 50). Hypandrium with long anteromedial rod-like apodeme, more robust than that of H. pollex. Gonostylus (Figs 49, 50, 56) dissimilar to those of all congeners, with pointed anteroventral corner and long anterior seta as in most relatives but ventrally externally with rounded and setose lobe and posteroventrally with distinctive, robust, dark, hooklike internal process. Aedeagal complex (Figs 52, 53, 55). Phallapodeme large and long, laterally flattened due to very large (high) dorsal and ventral keel. Aedeagus with peculiar distiphallus, most different from those of relatives because of small lateral lobes (Fig. 53), long ventromedial, posteriorly far projecting lobe (Fig. 52) and relatively short but broad funnel-shaped apex (Fig. 53); also phallophore distinctive, with short and more robust anterior part and small, slender, pointed posterior epiphallus-like projection (Fig. 52); postgonite relatively short and robust (Fig. 55), with expanded posterior lobe proximally and distinctly bent and pointed apex distally. Ejacapodeme reduced to very minute sclerite (see Fig. 53) hidden in posterior fissure of phallophore.
Discussion. Herniosina hamata sp. n. is a distinctive species, markedly different from all other species of the genus in the structures of the male and female terminalia. It is characterized by an interesting combination of plesiomorphic (e.g. reduced bulge on male S1+2; more setose male preabdominal sclerites; male S5 with a pair small medial projections; male cercus projecting in single process; gonostylus without dorsal internal projection; ejacapodeme minute but present; spermathecae simple, bulbous, without conical basal part) and distinctly derived autapomorphic features (gonostylus with hook-like posteroventral internal process; phallapodeme dorsoventrally dilated by both dorsal and ventral keel; distiphallus with long posteriorly projecting medial lobe; female postabdominal segments widened; female T8 dorsomedially divided; female S8 with "window"; cercus robust and with subapically inserted apical seta). H. hamata sp. n. most resembles H. pollex but the shared characters proved to be all plesiomorphic (male S1+2 with bulge reduced; male S5 with a pair of small medial projections; shorter male T5 and S8) and do not demonstrate their sister-species relationships. Indeed, the set of plesiomorphies of H. hamata sp. n. indicate that it could represent a sister-taxon to its four remaining congeners (see discussion of intrageneric relationships below).
Biology. Almost all type specimens were swept from low (and sparse) vegetation growing on layers of wet rotten plant debris. This is also true for the holotype being netted from "ruderal" vegetation on a pile of decayed leaves in a shaded ditch by a road. The longest series (9 specimens) was taken by netting over shooting plants on the wet shores of a small creek covered with decayed remnants of vegetation shortly after the snow melted (Fig. 66); also the habitat with decayed leaves close to a montane stream (Fig. 67) near the Caledonia Falls can be considered typical for the species. The altitudes of localities (all in Troodos Mts) range from about 900 m (Kakopetria env) to 1800 m (Troodos, close to Olympos Mt.) and all specimens were collected on 11-12 April.

5(2)
Preabdominal sterna sparsely setose (Fig. 41); cercus with medial process very long, robust, digitiform and projecting posteroventrally; its lateral process distally slender and laterally provided with a robust long seta arising on small lobe (Figs 37,38); gonostylus with lobe-like posteroventral part and internally with a small keel-like process (Fig. 39); phallophore anteriorly slender, ventromedial lobe of distiphallus simple (unmodified) and postgonite rather straight, with simple apex (Fig. 40)  Preabdominal sterna more densely setose (Fig. 51); cercus without medial process and its lateral process long, slender, apically somewhat dilated, with long seta arising more basally (Figs 49, 50); gonostylus with a robust posterior internal hook-like process directed ventrally and is posteroventral lobe smaller, knob-like (Figs 49,50,56); phallophore anteriorly thicker, ventromedial lobe of distiphallus projecting far posteriorly and of unusual shape (Fig. 52) and postgonite proximally dilated and with curved apex (Fig. 55)  T6 narrow, about as wide as T7 (Fig. 7); S10 divided into 2 small lateral sclerites (Figs 8,9); S8 small and usually with a small sclerotization in front of it in posteromedial emargination of S7 (Fig. 8)  T10 longer, elongately triangular (Fig. 32); S8 transversely suboval, with only 1 pair of setae (Fig. 33); cercus longer ( Fig. 34) .....H. horrida (Roháček, 1978) -T10 shorter, transversely triangular (Fig. 43); S8 more trapezoidal and with 1 pair of longer plus 1-2 pairs of short setae (Fig. 46); cercus shorter ( Fig. 45) Marshall, 1987), is a compact group of habitually very similar species differing mainly by the structures of the male and female terminalia. It is affiliated to the Limosina group of genera (Roháček 1982) and Apteromyia Vimmer, 1929 is considered its most closely allied genus (see also Marshall and Roháček 1982) based on the following synapomorphic characters: epandrium with a series of robust ventral lateral setae; distiphallus with unpaired ventromedial lobe projecting posteriorly; phallophore anteriorly slender and elongately projecting, movably attached to dorsal side of distiphallus; male cerci modified to compact processes below anal fissure. The relationships of species within the genus Herniosina can be hypothetized as follows. The set of plesiomorphies of H. hamata sp. n. (see in discussion under that species) indicate that it could represent a sister-taxon to the four remaining congeners which seem to form a monophyletic group supported by 5 synapomorphies: male preabdominal sclerites with setosity reduced; male cercus modified to 2 (lateral and medial) processes; gonostylus with dorsal internal projection; ejacapodeme absent; spermathecae pyriform, with distinct conical basal part. Herniosina pollex, having the male S1+2 with bulge reduced (a plesiomorphy shared with H. hamata) is obviously the sister-group to a cluster formed by H. horrida, H. erymantha sp. n. and H. bequaerti which possess the male S1+2 strongly protruding (bulging); moreover, this group also shares the very slender (in H. bequaerti secondarily shortened) dorsal internal projection of the gonostylus. Both these characters can be considered synapomorphies supporting relationships of these three species. Finally, H. horrida, with male S5 bearing a pair small posteromedial projections (a plesiomorphy shared with H. hamata sp. n. and H. pollex) can be postulated as sister-group to the remaining pair, H. erymantha sp. n. and H. bequaerti. Relationship of these sister-species is based on 3 synapomorphies: male S5 with posteromedial projections fused and prolonged to form a single, distally forked, process; medial apically pointed process of cercus small; funnel-shaped apex of distiphallus short and robust.
Habitat. Herniosina species seem to be originally terricolous phytosaprophagous flies associated with layers of decaying vegetation (leaf litter of broad-leaved trees in particular) in humid woodland habitats, as now known for H. hamata sp. n., H. horrida, H. erymantha, and partly also H. pollex. The latter species preferably lives on plant remnants in caves (but only close to their entrances) while H. bequaerti became entirely adapted to cavernicolous habitats developing in various rotting matter of plant and animal origin in caves (e.g. Czižek 1916;Duda 1918Duda , 1938 including their aphotic parts (Papp and Plachter 1976), cellars or mine galleries (see Roháček 1978Roháček , 1983 and burrows, runs and nests of various small mammals (Duda 1918;Richards 1930;Roháček 1978Roháček , 1983Rotheray 1991). Consequently, H. pollex, living in caves only temporarily as a component of the parietal fauna due to convenient conditions, can be classified only as a hemitroglophilous species while H. bequaerti, having the ability to develop deep in caves in complete darkness for generations (see Papp and Plachter 1976), is considered troglophilous despite the fact it can also develop in small subterranean habitats (burrows of mammals) or in cellars (Roháček 2014). The living habits of Herniosina species have partly reflected on their morphology, e.g. the somewhat reduced eyes, elongate arista, relatively strong sclerotization of body. They also are poor flyers, moving usually only by running and skipping on the substrate; hence they can be caught by sweeping with difficulty and only from very low vegetation or by netting over accumulated decayed plant remnants on the ground.
Biogeography. Two new species described above essentially contributed to the know ledge of the distribution of Herniosina. Particularly, its presence in the Eastern Mediter ranean was confirmed; previously there were a few records of unidentified Herniosina spp. from Israel (Papp and Roháček 1988: Mt. Hermon) and from Cyprus (Roháček 2004: Troodos Mts), but those of the latter are now attributed to H. hamata sp. n. Based on available data the distribution of the genus Herniosina ranges from Spain in the west to Russia (Kabardino-Balkariya) in the east and from Iceland and Fennoscandia in the north to Spain, Cyprus and Israel in the south Marshall et al. 2011). The most widespread species seems to be H. bequaerti being recorded from most of Europe (including Iceland) except for its southeastern parts (the absence of this species in caves of Balkan peninsula, cf. Séguy 1963, is particularly peculiar) and, surprisingly, H. pollex found besides Central Europe (Czech Republic, Slovakia) unexpectedly also in the Russian Caucasus (Kabardino-Balkariya). Other species may have a more restricted distribution, viz. H. horrida (Central Europe), H. erymantha (Greece: Pelopponese), H. hamata (Cyprus), Herniosina sp.cf. horrida (Israel). However, also these species can be more widely distributed considering the fact that the southern areas of W. Palaearctic are underinvestigated and that the terricolous or cavernicolous Hernisiona species are difficult to collect. supported by the Ministry of Culture of the Czech Republic by institutional financing of long-term conceptual development of research institution (the Silesian Museum, MK000100595), internal grants of the Silesian Museum Nos IGS201505/2015 and IGS201607/2016.