Hydraena (s.str.) dinarica, new species (Coleoptera: Hydraenidae) along with further records of Hydraena spp. from Durmitor National Park, Montenegro and comments on the DNA barcoding problem with the genus

Abstract Background Long-palped Water Beetles were collected during a taxon expedition in Montenegro which involved citizen scientists, students and taxonomists. The material was collected from springs, brooks, fens and the Tara River, at altitudes between 600 m and 1450 m above sea level, using fine-meshed hand-nets and by manual checking of submerged substrates. The morphological species delimitation was supplemented and congruent with mtDNA sequences mainly obtained in the field using the newly-developed MinION-based ONTrack pipeline. New information The new species Hydraena dinarica Freitag & de Vries, sp. n. from Durmitor Mt. is described, illustrated and compared in detail to closely-related congeners of the H. saga d'Orchymont, 1930/H. emarginata Rey, 1885 species complex. Five additional species and female specimens of two unidentified morphospecies of the genus were also recorded in the vicinity of Durmitor National Park. New records and the first DNA barcodes for Hydraena biltoni Jäch & Díaz, 2012 (endemic to Montenegro) and H. morio Kiesenwetter, 1849 are provided. Further records of H. nigrita Germar, 1824, H. minutissima Stephens, 1829, H. subintegra Ganglbauer, 1901 and females of two unidentified morphospecies are commented upon. The resulting inter- and intraspecific genetic distances and some observations of low or zero sequence divergence between recently-diverged species of Hydraena Kugelann, 1794 are briefly discussed.


Introduction
The Long-palped Water Beetles of the genus Hydraena, originally described by Kugelann (1794), represent the most speciose aquatic coleopteran genus. In tropical Asia, where the genus is still under-explored, a new species has even been discovered in the middle of a megacity (Freitag 2013). Nineteen Hydraena species are currently recorded from Montenegro. Many species of the genus are endemic to comparably small distribution ranges (Jäch et al. 2005), such as H. biltoni Díaz, 2012 andH. latebricola Jäch, 1986 in Montenegro (see Jäch 1986, Jäch andDíaz 2012). The Mediterranean area is particularly diverse and new species are still being discovered (e.g. Bilton 2013, Mičetić Stanković and Jäch 2012, Jäch and Díaz 2012. Therefore, the first author targeted Hydraena when instructing a citizen scientist project during a "taxon expedition" to Montenegro in 2019. The principles and great benefit of such initiatives are discussed by Schilthuizen et al. (2017) and Freitag et al. (2018). Through the enthusiastic support of citizen scientists, several aquatic habitats in and around Durmitor National Park, a UNESCO world heritage site, were sampled for aquatic beetles. The variety of collection sites included springs, creeks, lakes and fens in forests and alpine meadows, up to the stunning torrent of the Tara River in Europe's deepest gorge.
The subsequent identification of the collected material at the improvised field laboratory (Fig. 1, Suppl. material 2), set up in a holiday resort "Etno selo Šljeme" near the town of Žabljak, revealed eight Hydraena (s.str.) species, amongst which one new species was actually discovered. The new species was found at Skakala stream, a mountain creek flowing from Skakala waterfall into the periodically-inundated Sušica Lake on the northern slopes of Durmitor massif (Fig. 2).  Skakala stream at the northern slopes of Durmitor massif, type locality of Hydraena dinarica, sp. n.

Materials and methods
Specimens were collected in a microhabitat-specific approach (Freitag 2015) by disturbing the submerged substrates of the water body and collecting any floating specimens using fine-meshed hand-nets. Larger solid substrates (e.g. submerged wood) were taken off from the water and checked for specimens. The material was immediately preserved in vials with 96% ethanol, separately for each microhabitat assessed.
Pre-sorting and genus-level identification were performed by taxon expedition participants mentored by the first author ( Fig. 1), using taxonomic literature and dissection microscopes. Specimens of Hydraena were then dissected by the help of fine pincers and entomological pins to reveal their most diagnostic sexual characters (aedeagus, gonocoxite and tergite X) and compared with descriptions of Hydraena species of the Balkan Region. Their genitalia, temporarily mounted in lactic acid on microscopic slides, were examined under a Leica ICC50 HD compound microscope.
Detailed examination and digital imaging of dissected parts was done using an Olympus CX21 microscope equipped with a DinoEye Eyepiece camera. Habitus photographs were taken under a Zeiss Axio Zoom V 16 microscope with a Canon 5D Mark II SLR attached to the microscope. Images were captured at various focus planes and subsequently stacked using the Helicon Focus software. Genital drawings were compiled after their photographs by vector graphic tools in CorelDRAW v.10.0 software, but in direct comparison with the actual genitals mounted on slides.
After removal of diagnostic parts (glued on entomological cards), the entire remaining specimen of each initially-recognised morphospecies and some unidentifiable female specimens underwent DNA isolation, amplification, sequencing and processing of the 5′end of the mitochondrial cytochrome oxidase I (COI) gene as described in Maestri et al. (2019) (see https://github.com/MaestSi/ONTrack). The primer pair LCO1490 and HC02198 was used for PCR amplification. Library preparation for the MinION Oxford Nanopore NGS device was performed using SQK-LSK109 (Run1) and SQK-LSK108 (Run2) kits and, for each library, samples were pooled together after adding index sequences. The final libraries were loaded on a R9.4.1 Flongle flow-cell (Run1) and on a R9.4.1 MinION flowcell (Run2). Sequencing was carried out in the field using an off-line version of MinKNOW v1.6.11. The two sequencing runs were stopped after 5 and 17 hours and produced a total of 182,504 and 533,919 sequence reads, respectively. Sequence reads were base-called and demultiplexed using Guppy v.3.1.5 and accurate consensus sequences were generated using the ONTrack pipeline v.1.2.2 (Maestri et al. 2019). After our return from the field expedition, some additional sequences were generated by conventional Sanger sequencing by a commercial service using PCR products amplified as described above and following standard protocols. Forward and reverse Sanger reads were manually assembled into a consensus sequence using BIOEDIT version 7.2.5 (Hall 1999).
A statistical parsimony haplotype network was constructed by TCS1.21 (Clement et al. 2002), visualised using POPART (Leigh and Bryant 2015) and further edited in Adobe Illustrator 2020. The genetic sequence divergence analysis was performed in MEGA X using Kimura-2-parameter (K2P) model with the bootstrap method in 1000 replicates.
DNA sequences were submitted to International Nucleotide Sequence Database Collaboration (INSDC) through GenBank, as well as to the Barcode of Life Data System (BOLD) under project TXEX.
The type labels of the new species are literally quoted from the specimen's label under 'bibliographicCitation'. Back slashes indicate the next line in the label.

Distribution
Hydraena biltoni (Fig. 3A) is endemic to Montenegro. Previously, it was collected from the vicinity of Šavnik, about 20 km south of Žabljak and Biogradska Gora (Jäch and Díaz 2012).

Remarks
We provide the first standard barcode for the species. It varies by only 0.3% from that of the closest congener H. morio Kiesenwetter, 1849 (Suppl. material 1). As H. biltoni is extremely similar to the latter, which also occurs in the region, thorough examination of the aedeagus is required for proper identification.

Habitat
The specimens were collected in a very small creek, flowing through pine forest and a wet meadow. Bottom pebbles, mixed with CPOM, in moderately fast flowing, shallow portions of the creek, were their microhabitat.

Distribution
Hydraena minutissima (Fig. 3B), originally described from Great Britain (Stephens 1829), is widely distributed in southern, western and central Europe from Spain and Turkey in the south up to the British Isles (Jäch and Skale 2015).

Habitat
The species was collected in the shallow littoral of the Tara River with pebble deposits on bedrock. The microhabitat was not exposed to strong currents during the time of collection and filamentous algae were partly growing on the surrounding exposed rocks.

Distribution
Hydraena morio (Fig. 3C) is mainly distributed in eastern and central Europe in an area from Turkey to Germany, including the Balkan Region (Jäch and Skale 2015).

Remarks
We provide here the first COI 5′-end sequences (Folmer Region) of the species. See also remarks on H. biltoni.

Habitat
For notes on the habitat, see Hydraena minutissima.

Habitat
For notes on the habitat, see Hydraena minutissima.

Remarks
The samples cluster with Hydraena britteni Joy, 1907, originally described from England and Ireland (Joy 1907), vary only by ca. 0.3% and 2.5%, respectively, in their genetic distance from the latter. Based on the known distribution range of H. britteni, which does not include Montenegro and the genetic distance, it remains uncertain if either specimens are conspecific with the latter. Due to the lack of male specimens, we currently cannot identify these specimens with certainty.

Habitat
The specimens were collected from fen-like meadows, one (MNE10) densely vegetated with sedge and horsetail, the other (MNE13) additionally with limestone boulders and gravel densely covered with mosses. In both sites, a creek with clear brownish water, rich in humins, was passing the fens and provides continuous water inputs.

Distribution
The species is distributed in an area between the Adriatic and Black Seas, including the Dinaric Alps (Jäch and Skale 2015).

Remarks
The taxonomy of this species of the "Haenydra" lineage is not yet finally resolved. Three slightly varying morphs are recognised. Our specimens (Fig. 3E, F) belong morphologically and geographically to "Morph A" sensu Jäch and Díaz (2012). The standard DNA barcode of this morph which we are providing herein varies, in fact, by 0.5% from "Morph B" (Bulgaria).

Habitat
All specimens were collected from moderately fast flowing, shallow water, but on varying substrates, including submerged wood, grass bunches and pebble. Pronotum broadly subhexagonal, moderately wider than long; anterior and posterior margins slightly concave; anterior and posterior angles bluntly rounded, lateral rim denticulate, most conspicuous anteriorly; disc slightly convex; sagittal, anterior and posterior portions densely punctate; remaining disc portions moderately densely punctate; interstices glabrous; anterior and posterior sublateral foveae slightly impressed, rather inconspicuous; entire lateral portions slightly deflexed, rugulously bipunctate, partly microstriate.
Elytra elongate, almost parallel-sided apical 0.15-0.70; disc slightly vaulted, sublaterally more abruptly declivitous; elytral margin moderately explanate up to ca. apical 0.15. Elytra with six regularly arranged, not or slightly impressed rows of puncture striae between suture and disc declivity (approx. at the middle of shoulder) and ca. six additional, less regular puncture striae between disc declivity and elytral margin; punctures moderately large and moderately deeply impressed on anterior disc, gradually slightly decreasing in size and degree of impression towards apex and margin; intervals and interstices flat and glabrous; intervals smaller than puncture diameter anteriorly, larger in posterior and lateral portions; apical sutural teeth present or absent, apices separately rounded, sexually dimorphic (Fig. 5B, C).
Ventral side as in Fig. 5A. Mentum and submentum densely micropunctate. Genae and gula dominantly micropunctate, partly striate; posterior genal ridge distinct, glabrous. Hypomeron micropunctate to microreticulate. Prosternum densely micropunctate with hydrofuge micropubescence, with conspicuous median keel. Mesoventrite densely micropunctate with hydrofuge micropubescence, deeply impressed anterior to mesocoxae; impression transverse-arcuate; with pair of posteriad divergent glabrous streaks lateral to mesocoxae, across mesoventral impression (rather inconspicuous in some specimens); mesoventral disc and process convex. Metaventrite densely micropunctate with hydrofuge micropubescence, central disc shallowly impressed; Aedeagus (Fig. 6): Total length ca. 780 μm; main piece (630 μm long) with three long setae on inner (left) side and a very short one on outer (right) side; apex somewhat variable from convex to obliquely truncate, narrowest subapically; dorsal corner very slightly produced; main piece moderately slender, basal portion rectangularly bent from apical portion, gently narrowed apical 0.25 towards slender, subparallel apical portion; right margin (dorsal view) distinctly roundly produced at about mid-length; prebasal tooth short and blunt. Phallobase subsymmetrical in dorsal and ventral views. Distal lobe generally very similar of that in H. saga and related species, overall more stretched than compact; submembranous contorted distal portion relatively long and wide; opening funnel-like and apicad directed (like in a tuba), located at the most right (behind main piece and distal lobe trunk in dextrolateral view, Fig. 6B); most sclerotised enlarged distal portion with conically-pointed apex in dorsal and ventral views (Fig. 6A). Female tergite X (Fig. 5E), except for suboval shape, very similar to those of H. saga and related species (comp. Jäch and Díaz (2017): Figs. 7 and 12); apex widely rounded; disc with sub-basal squamose setae and with few trichoid setae; squamose setae comparably elongate and not conspicuously widened apically; subapical fringe admedially with dense fringe of vermiform setae of equal length which are slightly bent in apical half and with few long trichoid setae laterally.
Secondary sexual characters: Female elytral apices produced and separately gently rounded, not acuminate. All femora of male slightly more inflated. Male ventrite VI enlarged (Fig. 5A). Male mesotibia with a row of ca. ten denticles along proximal half of mesial face (Fig. 5G), in females only with setae (Fig. 5H). Male metatibia with fringe of long setae at inner face of posterior half (Fig. 5I), in females only with regular setae (Fig. 5J).
In comparison with all species mentioned above, Hydraena dinarica, sp. n. is unique in the tuba-like 180° bent hyaline distal tube of the aedeagus (Fig. 6)  dinarica, sp. n. is slightly larger (2.25-2.45 mm long) than the latter three species (1.95-2.30 mm long), the elytral disc appears slightly flatter and the elytral margin very slightly more explanate in H. dinarica, sp. n. The elytral apices are similar and within the observed variation range in the former species in both sexes. The new species also resembles H. samnitica of almost the same size (especially in the moderately large contorted aedeagal distal lobe), but it is externally distinguishable from H. samnitica by the explanate elytral margin extending almost up to the apex (vs. reaching apical 0.15; the apical area therefore appears more slender in H. dinarica, sp. n.).
On the other hand, H. dinarica, sp. n. seems genetically closest to H. alpicola, H. saga and the H. gracilis Germar, 1824 complex (as defined by Jäch (1995)), based on the DNA barcode (Fig. 7). H. saga occurs in the region (material at NMW; the closest known collection site is in Foča, Bosnia, less than 50 km away from the type locality of H. dinarica, sp. n.) and is also most similar. Therefore, the species can only reliably be identified by dissection of its aedeagus.
Males can be distinguished as stated above, while in females of H. dinarica, sp. n., the gonocoxite (Fig. 5F) is subtrapezoidal, with evenly round and expanded apical hyaline area (vs. subquadrate gonocoxite and with short apical area in H. saga;  : Fig. 11). Furthermore, the female tergite X in H. dinarica, sp. n. (Fig. 5E) is suboval, its basal portion expanded and the vermiform setae of the subapical fringe are bent (vs. subtriangular tergit X with short basal portion and with almost straight subapical vermiform setae in H. saga;  : Fig. 12).

Habitat
This species was collected from a mountain stream in a forested, undisturbed karst area at an altitude of about 1220 m a.s.l. (Fig. 2). During the time of collection, the river was only partly on the surface; the predominant flow was subsurface, causing low water temperatures. The well-shaded riverbed, including the water-bearing reaches, was densely covered with mosses. The specimens were collected from the upper interstitial of the bottom gravel (mesopsammon) in shallow, partly rapidly flowing water.

Distribution
So far only known from the type locality at the northern slopes of Durmitor Mt., Montenegro (Fig. 8).

Etymology
The species is named after the Dinaric Alps, or Dinarides, a karst mountain range where Durmitor Mt. and the type locality of the new species are situated. The epithet is used as an adjective meaning "of the Dinaric Alps". Map of Europe with the collection site of Hydraena dinarica sp. n. and the distribution of morphologically similar species of the "Haenydra" lineage defined as "Hydraena emarginata complex" by Trizzino et al. (2013a) and H. gracilis as genetically similar species with overlapping range.

DNA barcoding
The amplification of all sequences but one (H67), with LCO1490 & HC02198 primers and the applied protocols, was successful (Table 1).

Discussion
Due to the recent speciation, it is not surprising that we can barely delineate the species complex nor the individual new species by employing a 2% or 3% species delineation threshold of the mtDNA barcode as originally proposed (Hebert et al. 2003). This aspect is worth noting as it will be impossible to differentiate many young Hydraena species by the classical use of DNA barcodes.
Mitochondrial genes, in general and thus standard barcodes, often lag behind in terms of lineage sorting compared to nDNA involved in speciation (e.g. Monaghan et al. 2006, Balke et al. 2013 Figs. 1 and 2). The post-glacial, cold water habitat of H. dinarica suggests that it has evolved from a relictual population of a former cold-temperate/sub-polar climate-adapted ancestor related to H. saga, which had moved its distribution area southwards during the Pleistocene. Most probably, the new species could not survive in the lowlands of current Montenegro or even in regular surface waters of the same altitude, as reflected by its very restricted habitat and distribution. It can be assumed that the species withstands periods without surface flow (e.g. late summer) hidden in the hyporheic zone.
Hydraena dinarica clearly belongs to the " Haenydra" lineage (formerly regarded as subgenus by Berthélemy (1986) and Perkins (1997), a clade that was estimated to have occurred in the late Miocene (Tortonian), but however, is not satisfactorily resolved in terms of its phylogenetic position within the genus up to now (Trizzino et al. 2013b). Within the " Haenydra" lineage, H. dinarica, sp. n. belongs to the H. gracilis main clade of about 7-8 Ma age, which is morphologically evident in the articulated aedeagal distal lobe, as well as the low genetic distance from other representatives, based on the 657 bp barcode (Suppl. material 1). The very high morphological similarity, particularly in the articulate aedeagal distal lobe, gonocoxite and female tergite X, as well as the mtDNA data, clearly identify the new species as closely related to H. saga and H. alpicola. The terminology for the group, to which the latter two and other close congeners belong, is rather confusing and inconsistent. Most recently, it is referred to as H. saga complex sensu , previously as the H. emarginata / H. emarginata-saga complex in a wider sense (Trizzino et al. 2011a, Trizzino et al. 2013a including the H. belgica complex sensu Jäch and Díaz (2017), but all of them are not identical to the "H. saga complex" as mentioned by Ribera et al. (2011), referring only to the Iberian Peninsula representatives H. diazi and H. fosterorum. All of the latter are not well supported by DNA data, where H. saga and H. alpicola (of which mtDNA haplotypes have been reported to be sometimes identical (Trizzino et al. 2011b) usually cluster as a sister sub-clade to the H. graciles sub-clade, rather than with the other members of this group (current study, Ribera et al. 2011, Trizzino et al. 2013a). Irrespective of their morphological similarity, the polyphyletic appearance of the eastern (H. alpicola, H. dinarica, H. saga and possibly H. kahleni for which no genetic data are published) and western (all remaining species of the H. saga complex sensu Jäch and Díaz (2017)) sub-complexes are not congruent with either definition of such species complex. Redefining the species complexes appears necessary when more genetic information, including nuclear markers, becomes available.
Based on the DNA data available (e.g. current study, Abellán and Ribera 2017, Ribera et al. 2011, Trizzino et al. 2011b, it seems most reasonable to consider the eastern representatives and presumed sister group of the H. gracilis complex (Jäch 1995) as a newly-defined H. saga complex which, however, is not the purpose of this paper.
Many of these species are known to be highly endemic and all of them are young species, some of the closest relatives, such as H. saga and H. alpicola, have just split recently during the Pleistocene (about 50,000 years ago, based on data of Trizzino et al. (2013b)). A comparably young age can also be assumed for H. dinarica, sp. n. which coincides with the Würm glaciation that covered large parts of Durmitor massif in ice for the last time (Djurović 2009). Ribera et al. (2011) demonstrated a strong non-randomness of the geographic distribution of the species in some "Haenydra" clades, meaning phylogenetic distance is congruent to geographic distance of the related species (Fig. 8). This seems to be evident also for H. dinarica, sp. n. of which H. saga is the geographically closest member of the young species complex and is probably also the closest genetically. Therefore, post-glacial range fragmentation, due to the developing Mediterranean climate with a pronounced arid season, appears to be the most likely scenario for the evolution of H. dinarica, sp. n. The obvious ecological preference (or restriction) to its regionally-rare microhabitat with a partly subterranean flow of cold water supports this assumption.
Its rarity, its presumably very limited distribution range and its special habitat association suggest that the new species is particularly vulnerable to climate change and habitat destruction.