Two new species from the Hygrobates nigromaculatus-complex (Acariformes, Hydrachnidia, Hygrobatidae), based on morphological and molecular evidence

We analyse the taxonomic structure of the Hygrobates nigromaculatus-complex from the Balkan Peninsula. We describe two new species: Hygrobates lacrima Pešić sp. nov. (Montenegro) and H. limnocrenicus Pešić sp. nov. (Montenegro, North Macedonia). Although both species are morphologically similar, the average K2P-distance between DNA-barcode sequences from H. limnocrenicus sp. nov. and its closest relative H. setosus was 12.43% (SD = 1.47), and between H. lacrima sp. nov. and its closest relative H. nigromaculatus 15.87% (SD = 1.74). The new species exhibit distinct differences in terms of habitat preference: H. lacrima sp. nov. inhabits pools and shallow eddies along faster flowing waters, whereas H. limnocrenicus sp. nov. prefers deeper, fast flowing water, typically found in the outflow of a limnocrenic springs or lake outlets. The finding of these two new species suggests that efforts to investigate mites of the H. nigromaculatus-complex in the Balkans should be intensified.


Introduction
Water mites of the genus Hygrobates Koch, 1837 are often the most ubiquitous and usually the most abundant representatives of the group in different types of running and standing waters over the Palaearctic (Pešić et al. 2017). In terms of ecology, many species of this genus have been reported in literature to be present in both standing (lakes) and running (streams) water habitats. Such kind of bipolar habitat preference, for example has been reported for H. nigromaculatus Lebert, 1879 andH. longipalpis (Hermann, 1804), both often reported as common species in Europe living both in lakes and streams (see Martin et al. 2010 for a discussion). Recently, the status of these species has been questioned by the integrative studies using DNA barcodes, proving that lake and stream populations, indeed, represent morphologically and genetically distinct lineages (see Martin et al. 2010 for H. nigromaculatus;Pešić et al. 2019a for H. longipalpis).
Hygrobates nigromaculatus has been subject of controversial debate in taxonomy for a long time (see Martin et al. 2010 for an overview). The status of populations from Northern and Central Europe was resolved by a molecular study: using the DNA barcode region of the mitochondrial cytochrome c subunit I (COI) gene and the nuclear D2 region of 28S rRNA gene, the presence of two well-defined species could be revealed, H. setosus Besseling, 1942 living in streams, and H. nigromaculatus in lakes (Martin et al. 2010).
In this study we used morphological data and results of DNA-barcoding to analyse specimens of the H. nigromaculatus s.l. from the Balkans, with the aim to evaluate potentially cryptic species and establish the nigromaculatus species complex in the genus Hygrobates. As a result, two species new to science are described.

Materials and methods
Water mites were collected by hand netting, sorted live in the field, and immediately preserved in 96% ethanol. Specimens for molecular analysis were examined without dissecting under a compound microscope in ethanol, using a cavity well slide with a central depression. After DNA extraction, some specimens were dissected and slide mounted in Faure's medium.
Morphological nomenclature follows Pešić et al. (2017; for explanations concerning morphology and measurements of Hygrobates species see there Figs. 1B-D). The holotypes of the new species are deposited in Naturalis Biodiversity Center in Leiden (RMNH). DNA sequences prepared in the course of this study are published in BOLD with accession numbers indicated in Table 1.

Molecular analysis
Molecular analysis was conducted in the Department of Invertebrate Zoology and Hydrobiology, University of Łódź, Poland. For methods used for COI gene amplification and sequencing see Pešić et al. (2017). For this study, DNA was extracted from a total number of three specimens of genus Hygrobates from Montenegro and North Macedonia (Table 1).
Sequences were aligned by MUSCLE 3.8.425 algorithm as implemented in Geneious Prime 2020. 1.1 (Biomatters Ltd.). Phylogenetic tree for species delimitation was constructed  (Shimodaira and Hasegawa 1999) and Bayesian posterior probability (PP). Tree was edited in MEGA7 and further in Corel Draw X5. Pairwise distance calculations between nucleotide sequences were computed using Kimura's 2-parameter (K2P) distance model (Kimura, 1980) for all codon positions and transition/transversion ratio was calculated using MEGA7 (Kumar et al. 2016). Additionally, sequence data were analysed using the Automatic Barcode Gap Discovery (ABGD) method to delimit genetic clusters by detecting a significant gap in the pairwise distance distribution (Puillandre et al. 2012). We used the online ABGD version (http://wwwabi.snv.jussieu.fr/public/abgd/abgdweb.html) with default settings and K2P distance model.

Species delimitation using DNA-barcodes
The final alignment for species delimitation using COI sequence data comprised 603 nucleotide positions (nps) for 51 specimens including outgroups. The nucleotide sequences could be translated into amino acid sequences without any stop codons. In the dataset, 243 nps out of 603 were variable, and average transition to transversion ratio for all variable sites was 1.7. Molecular analysis shows that COI sequences from Hygrobates specimens collected in the Balkans form sister groups to the species previously known in the complex (Fig. 1). The sequence representing H. lacrima sp. nov. is reconstructed as a sister branch to the clade grouping COI sequences found in H. nigromaculatus, and two haplotypes found in H. limnocrenicus sp. nov. form a sister clade to the sequences found in H. setosus (Fig. 1); both relationships were recovered with strong support (0.90 PP, 88% SH and 0.99 PP, 96% SH, respectively).
The results of genetic distance analysis strongly supported the species status of the Hygrobates specimens collected in the Balkans. The genetic distance between the COI sequence of H. lacrima sp. nov. and its closest relative, H. nigromaculatus, was 15.87% (SD = 1.74) K2P, whereas the distance between H. limnocrenicus sp. nov. and H. setosus amounted to 12.43% (SD = 1.47) K2P. These distances were higher than the barcoding gap found by the ABGD method (3 to 11%) in the distances among all species belonging to the H. nigromaculatus-complex ( Fig. 2), which additionally supported the species-status of the two new clades. Diagnosis -Large in size (mL of Cx-I + gnathosoma ˃ 340, L genital plate ˃ 210, P-4 ˃ 170 μm); Cx-I+II apodemes protruding in both sexes, in females mediocaudal margins of Cx-IV with well-developed apodemes; anterior margin of male genital field with a bluntly pointed medial projection; L of IV-L-6 proximoventral seta ♂ ˂ 20, ♀ ˂ 35 μm.
Male -Anterior margin of genital field convex, with a small bluntly pointed medial projection, posterior margin indented, with a small central protrusion not extending beyond posterior genital plate margin (Fig. 5A).
Remarks -Based on molecular analysis, Hygrobates lacrima sp. nov. is closely related to H. nigromaculatus. The distance between these two species was 15.87% (SD = 1.74) K2P.
In  (2017) showed different types of male genital plates for H. nigromaculatus, we cannot exclude that he dealt with several species. Therefore, we think that the differences between the male genital plates of H. lacrima and other species as given in the key are justified.
Due to its larger size and habitat preference for running waters, the new species is similar to H. limnocrenicus sp. nov. (see there for further discussion).
Distribution -Montenegro; know from two localities along the middle course of the Tara River (Figs. 9A-C). These two sites were subject of a monthly monitoring survey in 2019 on the river biota of the Tara River during the Bar-Boljare highway development activities (see Pešić et al. 2020a). At the Mataševo site, only one specimen (which was successfully barcoded) was collected, probably due to the negative ecological impact associated with highway development activities in the immediate vicinity (for the controversy surrounding highway construction over the Tara River see Pešić et al. 2020b). At the Trebaljevo site located downstream, the negative impact was less pronounced -here, three individuals were collected. Diagnosis -Large in size (mL of Cx-I + gnathosoma ˃ 350, L genital plate ˃ 200, P-4 ˃ 170 μm); Cx-I+II apodemes moderately protruding, mediocaudal margins of Cx-IV in females without prominent apodemes; anterior margin of male genital field with a knob-shaped medial projection; L of IV-L-6 proximoventral seta ♂ ˃ 40, ♀ ˃ 35; running waters.
Male -Anterior margin of genital field convex, with a small knob-shaped medial projection, posterior margin indented, with a rounded central projection not extending beyond posterior genital plate margin (Figs. 5E).
Etymology -The species is named after its dominant occurrence in limnocrene springs. Remarks -Due to its larger dimensions (and also regarding habitat preference), the new species is similar to H. lacrima sp. nov. Females of the latter species differ from H. limnocrenicus sp. nov. in having well-developed apodemes at mediocaudal margins of Cx-IV (compare Fig. 3A with Fig. 6A), males in the medial projection of the genital field bluntlypointed (knob-shaped in H. limnocrenicus sp. nov.; compare Fig. 5A and 5E), both sexes bear a shorter proximoventral seta on IV-L-6 (L H. lacrima vs. H. limnocrenicus, ♂♂: ˂ 20 vs. ˃ 40, ♀♀: ˂ 35 vs. ˃ 35 μm).
Analysis of COI sequences suggests that H. limnocrenicus sp. nov. is most closely related to H. setosus Besseling, 1942, which has also a preference for running water habitats. However, the latter species, as well as H. lacrima sp. nov., was mostly found in pools of running waters, while H. limnocrenicus sp. nov. prefers deeper, fast flowing water, typically being found in the outflow of limnocrenic springs or lake outlets (Figs. 9D-F). The COI divergence between H. setosus and H. limnocrenicus sp. nov. was 12.43% (SD = 1.47). K2P indicating a long independent history of these two species. The studied specimens of H. setosus (in parentheses) differ in both sexes by a shorter proximoventral seta on IV-L-6 (L H. setosus vs. H. limnocrenicus, ♂♂: ˂ 30 vs. ˃ 40, ♀♀: ˂ 20 vs. ˃ 35 μm) and in males by the comparatively less wide genital field (L/W ratio ˃ 0.8), a smaller dimensions of acetabula (Ac-3 ˂ 105 μm), and the medial projection of the genital field less pronounced, typically with irregular margin of a secondary sclerotization (see Figs. 5G-H).
Biology -Morphological and genetical analysis of populations from the Vitoja springs situated at the northeastern shore of the Lake Skadar indicates that the oviposition data published by Bańkowska et al. (2016) for H. setosus in fact refers to H. limnocrenicus sp. nov. The highest number of females laying eggs was found in springs, with an average number of eggs per female of 58.2±30.6. A lower number of females laying eggs was noted in rivers, but here the average number of eggs per female (80±24.5) was higher (Bańkowska et al. 2016). A statistically significant difference in number of laid eggs between May and October was found. The average time of hatching was 14.1 ± 5.29 days.
Distribution -Montenegro and North Macedonia (Fig. 8). Details of distribution are unknown due to the previous confusion with H. setosus, but the species is obviously widespread in the Balkans. It is likely that the records of the latter species from the limnocrene springs of the Mediterranean region of the Balkan refers to H. limnocrenicus sp. nov. For example, Pozojević et al. (2019) reported occurence of H. setosus in the limnocrene spring Modro Oko in Southern Dalmatia, in a low abundance not exceeding 4 individuals per square meter.

Figure 9
Sampling sites: A-B -River Tara near Trebaljevo; C -River Tara near Mateševo; D -the "Kraljičino Oko" (the Queen's Eye) spring near Podgorica; E -River Crni Drim near Struga; F -the Vitoja spring located on the north-eastern shore of Lake Skadar flooded in the winter and spring months.
-Male: genital field comparatively less wide (L/W ratio ˃ 0.8), acetabula smaller in dimension, Ac-3 ˂ 110 μm, anteromedial projection of genital field less pronounced, often with irregular border of secondary sclerotization (Figs. 5G-H); IV-L-6 proximoventral seta L: From H. setosus and two new species described from the Balkans, H. nigromaculatsus differs in smaller size and its habitat preference for standing waters, often preferring surge shores of the littoral zone of lakes (Martin 1996, Martin et al. 2010. Hygrobates setosus is exclusively found in pools of streams with slow current (Martin et al. 2010). Similar to H. setosus, H. lacrima sp. nov. inhabits pools and shallow eddies along faster flowing waters such as the Tara River. Finally, H. limnocrenicus sp. nov. prefers deeper, fast flowing water, typically found in the outflow of a limnocrenic springs or a lake outlets.
During our survey, an abundant population of H. limnocrenicus sp. nov. was found in the Vitoja springs, located at the shore of Skadar Lake, and falling below the lake water level during high water periods in winter and early spring (Pešić et al. 2018b; see Fig. 9F). Probably, the proximity of the lake provides optimal conditions to sustain a large population of H. limnocrenicus sp. nov. compared with isolated water bodies where it was found in a low abundance only (regardless of the effort made to sample more material, only a few specimens were collected in the outflows of the Mareza and Kraljičino Oko (Queen's Eye) springs). It is worth mentioning that H. limnocrenicus sp. nov. was found also in the Crni Drim (Black Drin) River near the exit from the Ohrid Lake which shares the same watershed as the Skadar Lake. The latter finding supports the congruence of the distribution of this species in the Adriatic Sea watershed.
As demonstrated by Martin and Davids (2002), differences in life-cycle strategies may contribute to species discrimination within the H. nigromaculatus-complex. The streamdwelling populations of H. setosus have a parasitic larva, whereas the lake-living populations of H. nigromaculatus have reduced larval parasitism (Martin and Davids 2002). The question if parasitism eventually went lost in a common ancestor of H. lacrima sp. nov. and the H. nigromaculatus merits particular attention. Martin et al. (2010) hypothesized that H. nigromaculatus secondarily lost parasitic larvae and probably originated from a stream-living Hygrobates species with a larval stage able to disperse. Both new species have a relatively large body size which may be a first hint to a parasitic life style. Often, non-parasitic lineages are able to afford to be smaller by producing less abundant, but larger eggs (Smith 1998).
Although our study is based on a relatively small number of individuals that could be sequenced, it resulted in the discovery of two new species and the definition of a H. nigromaculatus species complex. Our results suggest that efforts to investigates mites of this complex should be intensified all over Europe.