Coelosynapha, a new genus of the subfamily Gnoristinae (Diptera: Mycetophilidae) with a circumpolar, Holarctic distribution

Abstract Background The subfamily Gnoristinae is one of the most diverse and taxonomically difficult subfamilies of Mycetophilidae, with new species and genera being described almost every year from various parts of the world. Through inventories of fungus gnats in the Nordic Region and Russia, a genus and species new to science was discovered, yet with links back to an illustration made by the late French entomologist Loïc Matile in the 1980s. DNA barcoding aligned it with yet another species new to science, distributed across Canada and documented through The Barcode of Life Data System (BOLD) by Paul D. N. Hebert and colleagues at the BOLD team. New information The new Holarctic genus, Coelosynapha gen. n. is described, consisting of two new species, the Palaearctic Coelosynapha loici sp. n. and the Nearctic Coelosynapha heberti sp. n. DNA-barcodes assign the two new species to distinctly separated (8.27% p-distance) Barcode Index Numbers (BINs) which are most closely aligned to unidentified species of Mycetophilidae from South Australia and Costa Rica on BOLD. The new genus shows morphological characteristics in between the two Holarctic genera Coelosia Winnertz, 1864 and Synapha Meigen, 1818 and further shows affinity to the southern continents genus Austrosynapha Tonnoir, 1929. The Palaearctic Coelosynapha loici sp. n., for which habitat requirements are best documented, is largely restricted to pristine, old-growth conifer (mostly spruce, Picea abies ssp. obovata) forests within the boreal vegetation zone, although it is also recorded from hummock tundra along the Anadyr River in Far East Russia.


Introduction
According to Fontaine et al. (2012), the average shelf life between discovery and description of a new species is 21 years across different taxa and the longest average shelf life is documented for terrestrial animals discovered by professional taxonomists living in rich countries. Sitting on huge collections with hundreds of potentially new species to science in our Nordic museum collections, the authors of this paper are guilty of this misconduct. For higher taxa, like describing a new genus, we suspect the average time can be even longer, given the uncertainty in relating it to known taxa often coupled with the scarcity of specimens of high quality. Still, discovering and describing new genera from the comparatively well-known insect fauna of the Nordic Region (see, for example, Ronquist et al. 2020) is not commonplace anymore.
The current case concerns an undescribed genus known from the high north of Fennoscandia, from the Altai mountains and Far East Chukotka in Russia and from across southern Canada, i.e. a circumpolar taxon. Some 25 years ago, Geir E. E. Søli at the Natural History Museum in Oslo was made aware of an enigmatic species belonging to the subfamily Gnoristinae of the family Mycetophilidae when the late French entomologist Loïc Matile (1938( -2000( , see Daugeron et al. 2002) sent him illustrations (Fig. 1) and a brief onepage description of a potential new genus, based on a single specimen collected in Finland. On the top of the description page, Loïc Matile had noted by hand: "Had this for years… What do you think? LM". Due to the scarce material, the case was shelved until the species was collected again, in 2009, this time also in Finnish Lapland by one of the authors (JS). Specimens were this time sent to a specialist in the USA, but again shelved without further progress. Further materials appeared from across Russia and Northern Norway. Meanwhile, after gathering at the 8 International Congress of Dipterology in Germany in 2014, the authors of this paper started tighter cooperation on the discovery of new and problematic species of fungus gnats in the Nordic Region through a forum page on the Vibrant ScratchPad site "Fungus Gnats Online" (Fungus_Gnats_Online 2020). We all run national projects to record and describe the Nordic and Russian fauna of fungus gnats (see, for example, Jakovlev et al. 2014, Kjaerandsen et al. 2007b, Polevoi 2010, Polevoi and Barkalov 2017, Salmela and Kolcsár 2017Kjaerandsen et al. 2007a, Kjaerandsen and Jordal 2007 and we are heavily engaged in DNA barcoding through the NorBOL (NorBOL 2020, Kjaerandsen 2017) and FinBOL (FinBOL 2020) initiatives. Eventually, specimens of the enigmatic new species were obtained from several, mainly old-growth, coniferous sites across the Palaearctic Taiga, ranging from Reisa National Park in Norway in the west all the way to Chukotka in the Far East of Russia. Some 25 year ago, the late French entomologist Loïc Matile  prepared these illustrations and a brief one-page description of a potential new genus based on a single specimen collected in Finland (A = ventral view, B = dorsal view.) aligned them as the nearest neighbour to another unidentified and very similar species, sampled across southern Canada and barcoded through the Canadian Barcode of Life initiative (Hebert et al. 2016). Upon request to the Centre for Biodiversity Genomics, we were kindly offered a loan of the voucher material of this species, which, together with our Palaearctic material, are argued here to represent a new genus of the subfamily Gnoristinae, family Mycetophilidae.
The subfamily Gnoristinae appears to be amongst the most difficult branches of the Mycetophilidae to classify. Phylogenies (e.g. Kaspřák et al. 2018) are still rendering it paraphyletic with respect to Mycetophilinae and, according to Kaspřák et al. (2018), the Gnoristinae is one of the most diverse and taxonomically difficult subfamilies of Mycetophilidae, with new species and genera being described almost every year. Highly variable taxa have led to numerous small genera with few species being segregated, as well as species-rich, polyphyletic "trash bin" genera like Dziedzickia Johannsen, 1909. Within these genera, the variation in the classical Meigen-Winnertzian character system, which is largely based on wing venation, tends to break down, especially when tropical taxa are considered (see further discussion in Ševčík and Kjaerandsen 2012).

Specimen preparation and storage
The studied specimens were collected over a period from 2009 to 2016 from 12 localities in North America and Eurasia (Fig. 2). Being initially stored in 70-95% ethanol, they were partly dried through baths of hexamethyldisilazane (HMDS, see Brown 1993) and pinned during the study. Terminalia were detached from the abdomen and treated by standard methods (macerated either in warm lactic acid or in a solution of potassium hydroxide (KOH), cleaned in distilled water and neutralised in acetic acid). After detailed study and imaging, the terminalia were placed into microvials with glycerine and pinned together with the rest of the specimen. The poor (fragmented) quality of the voucher materials borrowed from the Centre for Biodiversity Genomics did not allow for dry pinning, so the primary types from Canada were transferred from alcohol to glycerine in microvials on pins. Materials are deposited in the collections of the following institutions: Centre for Biodiversity Genomics, University of Guelph, Canada; California State Collection of Arthropods, Sacramento, California, USA; Tromsø University Museum, Tromsø, Norway; Regional Museum of Lapland, Rovaniemi, Finland; Siberian Zoological Museum, Novosibirsk, Russia; Tomsk State University, Tomsk, Russia; Forest Research Institute, Petrozavodsk, Russia.

DNA barcoding
The 658 bp fragment of the mitochondrial protein-encoding cytochrome c oxidase subunit I (COI) was sequenced from a total of 10 Coelosynapha loici sp. n. specimens and five C. heberti sp. n. specimens. One leg from each fresh specimen was sent to the Canadian Centre for DNA barcoding, BIO (Guelph, Ontario, Canada), for DNA extraction and bi-directional Sanger sequencing as a part of the Norwegian Barcode of Life (NorBOL) (see Kjaerandsen 2017) and Finnish Barcode of Life (FinBOL) initiatives, both branches of the International Barcode of Life project (iBOL). The new sequences are available from The Barcode of Life Data System (BOLD) and also as supplementary material (Suppl. material 1).

Illustrations
A Leica MC170HD microscope camera, mounted on a Leica M205C stereomicroscope, was used to capture images of whole specimens and of detached terminalia macerated in hot lactic acid and stored in glycerine. Stacked images, merged for extended focus applying the Helicon Focus software, were subsequently moderately photo-shopped into illustrative plates. Digital illustrations (Fig. 10) were made with Inkscape vector drawing editor (http://inkscape.org).

Terminology
The general terminology of body, wings and terminalia follows Søli (2017) with a few additional, more specific terms from Söli (1997). The circumpolar distribution of Coelosynapha gen. n. visualised on Google Earth. Green plots across the Palaearctic region display records of Coelosynapha loici sp. n., yellow plots across the Nearctic region display records of Coelosynapha heberti sp. n.

Description
A Gnoristinae genus, as presently known with moderately slender and quite small, down to 3 mm body length, species (Figs 3,5). Colouration uniformly brown on head and body, darker preterminal segments, mostly yellow on legs and terminalia. Head ( Fig. 4) round, eyes kidney-shaped with tendency of dorsal eye-bridge expansion (like in Synapha Meigen, 1818, unlike in Coelosia Winnertz, 1864and Austrosynapha Tonnoir, 1929, interommatidia pubescent. Antenna moderately slender, with 16 segments, large, semi-globular pedicel and flagellar segments 2-3 times as long as wide (shorter in Synapha, distinctly longer in Coelosia and Austrosynapha). Mouth parts average, with five, gradually longer palpal segments, no clear sensory pit discernible in third segment (without slide mounting). Clypeus bud-shaped, shorter than face. Three ocelli in a near straight line (Fig. 4b, c), lateral ocellus more than two times its diameter from eye. Antepronotum with pair of strong antepronotal setae arching over the head (Fig. 4a, c). Mesonotum with setae in acrostichal and dorsocentral rows, devoid of setulae in between but rich in setae laterally (Fig. 4c). Meso-and metapleurites all without setae (Figs 4a, 5). Wings (Figs 6, 7) hyaline, unpatterned, wing membrane with irregularly arranged microtrichia. Costa produced more than half way between R and M , subcosta long, ending in C proximal to crossvein Rs ( Fig.  6a), usually without, occasionally with crossvein sc-r (Fig. 7d). Radial sector variable, usually with R present (Fig. 6d), sometimes with R and R separate (Fig. 6b), occasionally with only crossvein Rs (Fig. 7b). Anterior fork with stem more than 2× longer than r-m. Posterior fork short, rather widely divergent. All veins anterior of iCu with setae on dorsal surface except for basal transversal crossvein tb and M-stem. Legs with irregularly arranged setulae. Fore tarsus subequal in length to fore tibia. No sense organ discernible on mid tibia.

Diagnosis
A Gnoristinae genus similar to Austrosynapha Tonnoir, 1929, Coelosia Winnertz, 1864 and Synapha Meigen, 1818 in general appearance, but with very characteristic and unique male terminalia with three blunt, digitate processes apically on the gonostyles (Figs 9a, b, 10a, b). It can be separated from the three genera by the wing venation having the combination of 1) extension of C long, ending more than half way between R and M (Fig. 6a, like in Coelosia, shorter in Austrosynapha and Synapha); 2) Sc ending in C at level of Rs (Fig. 6a, shorter in Austrosynapha which is variable for this character), usually without, but occasionally with sc-r present (Fig. 7d, always absent in Coelosia); 3) anterior fork petiole more than 2× length of crossvein r-m (Fig. 6a, like in most Austrosynapha and all Synapha, unlike in Coelosia); 4) short and wide posterior fork (Fig. 6a, like in all Coelosia and some Austrosynapha, unlike in Synapha and some Austrosynapha). Variation in wing venation of males of Coelosynapha loici sp. n. Note the substantial variation in the radial sector ranging from R and R being absent to separated to a variable degree of fusion into R . Variation in these characters are also apparent between left and right side in some specimens. Abbreviations: A = anal veins (numbered); CuA = anterior branch of cubital vein; CuP = posterior branch of cubital vein; h = humeral vein; iCu = intercubital fold (not a vein); M = medial veins (numbered); R = radial veins (numbered); r-m = radial-medial crossvein; Rs = radial sector; sc = subcostal vein; tb = basal transversal.

Etymology
The generic name is feminine gender and put together by the two genus names Coelosia Winnertz, 1864 and Synapha Meigen, 1818, indicating the affinity to and intermediate position between those two genera.

Distribution
Records of the new genus display a circumpolar distribution pattern from Fennoscandia to Far East Russia in the Palaearctic Region and across Canada in the Nearctic Region (Fig. 2).
Variation in wing venation of Coelosynapha heberti sp. n. Note the substantial variation in the radial sector ranging from R being absent or present. Note also the appearance of sc-r in one female wing, while all the others are lacking this crossvein. The most prominent difference in the wings of the two species appears to be the relatively longer posterior fork with basally less curved M in Coelosynapha heberti sp. n. Abbreviations: R = radial veins (numbered); scr = subcostal-radial crossvein. a: Holotype, male, whole wing. b: Holotype, male, details of radial sector. c: Paratype, female, whole wing. d: Paratype, female, details of radial sector.
Thorax (Fig. 4a, c). Scutum usually dark-brown (with yellow humeral area in Chukotka specimen), thinly dusted, with yellow setulae. Pleural sclerites glabrous, dark brown, Strong antepronotal seta about as long as head. Scutellum with two strong setae and four small setae.
Wings (Fig. 6). Wing length 2.6-2.9 mm. Sc ending in C, cross-vein sc-r absent. C produced beyond apex of R5 for about 2/3 of the distance between R and M , R present, forming rectangular cell, occasionally separate veins R and R developed. Anterior fork long, divergent. Posterior fork short, its long stem is forked well beyond point of furcation of anterior fork. Veins M and CuA rather widely divergent. Sc with 3-5 macrotrichia on apical half. Basal transversal crossvein (tb) and M-stem bare, all other veins anterior of iCu fold setose. Halters pale yellowish.
Legs. Legs yellow with a brown tinge on trochanter, tibia and tarsi. Fore tarsus 1 slightly longer than fore tibia (ratio 1.06 in holotype).

Etymology
Coelosynapha loici sp. n. is named in honour of the late French entomologist Loïc Matile  who first studied the new species and illustrated its terminalia (Fig.  1) way back in the mid 1980s. It is further an expression of the importance of classical morphological studies including detailed illustrations to be retained in the new integrative science of taxonomy.

Distribution
The new species has a wide Palaearctic range in boreo-mountainous localities, it has been collected from the High North boreal forest of Fennoscandia (Norway, Finland, NW Russia), from a high mountain site in Asian part of Russia and from hummock tundra in Far East Russia, at a total of seven localities (Fig. 2).

Ecology
Six of the collecting sites are within the boreal vegetation zone while the record from Chukotka is from the low arctic ecoregion. The European sites, one in Norway, three in Finland, one in Russia, are all northern boreal, close to the range limit of spruce (Picea abies ssp. obovata), lying some 140-180 km north of the arctic circle. All European collecting sites are pristine, boreal old-growth forests, dominated by spruce or occasionally pine (Reisadalen NP) and with scattered deciduous trees (birch, Betula pubescens and goat willow, Salix caprea). The ground layer is characterised by mosses, especially Pleurozium schreberi and bilberry ( Vaccinium myrtillus). Malaise traps on these sites have been set in the vicinity of springs or cold headwater streams. The specimens from Altai were collected at an altitude of 2144 m above sea level in steppe type Larix forest. The Chukotka specimen was collected with a yellow pan trap set in hummock tundra along the Anadyr river at 5-10 m above sea level.
Wings (Fig. 7a, b). Wing length 2.6 mm. Crossveins sc-r and R absent (in holotype, Fig. 7b). Posterior fork longer, its stem forking opposite of anterior fork (Fig. 7a). Veins M and CuA branching with acute angle and less divergent at base of fork. Sc with 6 macrotrichia on apical half.

Etymology
Coelosynapha heberti sp. n. is named in honour of Paul D. N. Hebert, "the father" of DNA barcoding who also led the project of barcoding the insects of Canada (Hebert et al. 2016) from which the new species was located through its barcode similarity with Coelosynapha loici sp. n. It is further an expression of the importance of DNA barcoding in the new integrative science of taxonomy.

Distribution
The new species has a wide Nearctic range across Canada (Fig. 2).

Section of the generic key to separate Coelosynapha and Coelosia
Couplet 1  2 Stem of anterior fork more than 2x longer than crossvein r-m, R present or absent, crossvein sc-r present or absent Coelosynapha gen. n.
-Stem of anterior fork at most slightly longer than crossvein r-m, R always absent, Crossvein sc-r always absent Coelosia Winnertz, 1864

Discussion
The enormous success of DNA barcoding now has accumulated a substantial amount of sequenced insects on BOLD, very useful for new and integrative taxonomic studies. More than 65,000 specimens belonging to the family Mycetophilidae have been successfully sequenced and, of them, some 10,000 are assigned to the subfamily Gnoristinae. Some 1100 identified Mycetophilidae species have public barcodes although more than 2400 different BINs are assigned, thus indicating that the majority of the species still remains unidentified beyond the (sub)family level on BOLD. A weakness with the BOLD initiative may be that several of the typically well-funded, large scale DNA barcoding projects, undertaken so far, did not have a focus on, nor adequate resources allocated to, securing high quality morphological identification of the vouchers for the accumulated barcodes. Unfortunately, this critical endeavour of the BOLD archive is largely left to the under-funded and scarce taxonomic expertise to engage in "post-sequence" at will.
In the Nordic region, however, strong ties between The Norwegian and Swedish Biodiversity Information Centres, including their Taxonomy Initiatives and NorBOL and FinBOL, are ensuring that the best taxonomic expertise is building up the reference library of the local fauna on the BOLD archive. Hence, the vast majority of some 6500 DNA barcoded fungus gnats (Sciaroidea) from the Nordic region have been identified to species level upon submission and the reference library is profoundly and repeatedly qualitychecked and curated after barcodes and BINs are assigned. This has resulted in a high quality reference library, now covering about 90% of the known fauna and more than 100 additional species considered to be new to science (Kjaerandsen 2020, Kjaerandsen 2017). Hence, when the sequences of Coelosynapha gen. n. most closely resembles exotic, unidentified species of Mycetophilidae from South Australia and Costa Rica (Fig. 14a), this is a strong indication of a genus not previously known from the region. When we restrict the dataset to the 6500 sequences, representing nearly all genera of Nordic Sciaroidea, Coelosynapha gen. n. is most closely aligned with the genus Palaeodocosia Meunier, 1904 (Fig. 14b), while species of both Synapha and Coelosia appear more distant.
Coelosynapha loici sp. n. is assigned to the BIN BOLD:ADD0785, consisting of 10 members with a maximum within-species genetic distance of 0.72%. Coelosynapha heberti sp. n. with five members assigned to BIN BOLD:ACI5160 likewise has a maximum withinspecies genetic distance of 0.72%. The reciprocal nearest-neighbour distance between the two is 8.27-8.66% (depending on direction). A barcode gap analysis of the distance between Coelosynapha loici sp. n. and close genera reveals the closest Coelosia being Coelosia truncata Lundström, 1909 at 13.67%, the closest Synapha being Synapha fasciata Meigen, 1818 at 15.36% and the sole determined Austrosynapha sp. on BOLD (sp. JSGS1, mined from GenBank) at 16.79% distance.
The variation seen in the radial sector of the wing in species of Coelosynapha gen. n. is not unique. Several Gnoristinae genera show variation in exactly this character, including Grzegorzekia Edwards, 1941, Speolepta Edwards, 1925, Synapha and Syntemna Winnertz, 1864, to a point where the instability is almost a characteristic of parts of the subfamily. None of these genera, however, appears to be very closely related to Coelosynapha gen. n. as revealed by COI haplotype similarities.
There are indications that both Synapha and Austrosynapha, as presently defined, are polyphyletic genera in need of revision, but it is beyond the scope of this paper to address these issues pending a more thorough revision of the entire subfamily. Here we have added another genus that, if included in either of them or in Coelosia, would likely render them even more para-or polyphyletic. We hope that describing Coelosynapha gen. n. will ID-tree (Kimura-2-distance) obtained from BOLD of DNA barcodes of Coelosynapha gen. n. and close allies.