A new species of mayflies Bungona (Chopralla) pontica sp. n. was collected in northern Turkey in 2011 (Fig. 8). Herein, we describe the new taxon including both, morphological and molecular techniques. To analyze the affinities of the new species based on the molecular data, we built a dataset comprising sequences of several mayfly taxa, as summarized in Table 1.

In the selection of taxa, we wanted to include all major clades within the Cloeodes-complex, as inferred based on morphological data by Salles et al. (2016). Specifically, the sampling included the genus Potamocloeon Gillies, 1990 with the subgenera Potamocloeon and Aquaediva Salles et al. 2016, which together constitute a monophyletic group sister to all the rest of the Cloeodes-complex. We also included the genus Crassolus Salles et al. 2016, forming a sister lineage to the clade comprising the genera Cloeodes and Bungona. These latter two taxa were also included in the analysis, in case of the genus Bungona, all three subgenera (Bungona, Centroptella Braasch & Soldán, 1980 and Chopralla Waltz & McCafferty, 1987) were sampled. Therefore, we achieved a comprehensive set of sequences from all major phylogenetic lineages within the Cloeodes-complex. We also included some taxa outside of the complex, considered related to the ingroup based on previous morphological investigations (Salles et al. 2016). These encompass the genera Crassabwa Lugo-Ortiz & McCafferty, 1996, Cheleocloeon Wuillot & Gillies, 1993, Dabulamanzia Lugo-Ortiz & McCafferty, 1996 and Centroptilum Eaton, 1869.

A part of the sequences used for the analysis were obtained from the public databases, the rest was newly sequenced within the framework of the present study (see Table 1 for details including sources of data and accession numbers). We also searched in BOLD and GenBank databases for other existing similar sequences. However, no sequences more similar than 84% (and therefore potentially informative for the analysis) were found except those already included in the dataset.

Voucher specimens of the new species, as specified in Results, is deposited in the collection of the Biology Centre of the Czech Academy of Sciences, Institute of Entomology, České Budějovice, Czech Republic (IECAS: holotype larva, three paratype larvae) and in the Musée cantonal de zoologie, Lausanne, Switzerland (MZL: two paratype larvae).

Two paratype specimens of Bungona (Chopralla) pontica sp. n. were mounted on slides, using HydroMatrix (MicroTech Lab, Graz, Austria), Canada balsam, or Euparal. Drawings were made using an Olympus SZX7 stereomicroscope and an Olympus BX41 microscope, both equipped with a drawing attachment (camera lucida). Photographs were made using a Leica M205 C stereomicroscope with a Canon EOS 1200D digital camera. All photographs were subsequently enhanced using Adobe Photoshop CS5. For scanning electron microscopy (SEM), the samples in 96% ethanol were transferred in three mixtures of graded series ethanol : acetone (2:1, 1:1 and 1:2) with 15 minutes in each mixture and dried by means of a critical point dryer CPD 2 (Pelco TM). The samples were mounted on an aluminium target by means of carbon tape, coated with gold using a sputter coater Baltec SCD 050. Finally, the samples were examined in a FE-SEM JSM 7401-F scanning electron microscope (JEOL Ltd, Tokyo, Japan) at low accelerating voltage of 4 kV using GB-low mode. The accomplished topographic images provided a resolution of 1.5 nm by an efficient semi-in-lens detector. The morphological terminology follows Salles et al. (2016).

DNA extraction, amplification and sequencing. Genomic DNA was extracted from whole specimens or legs using the NucleoSpin 96 (Macherey-Nagel, Düren, Germany) tissue kit, the BioSprint 96 extraction robot (Qiagen), or the Wizard SV Genomic DNA Purification System Kit (Promega, California, USA), following the manufacturer's protocol. We analyzed the 658 bp fragment of the mitochondrial DNA barcoding gene cytochrome c oxidase subunit 1 (cox1). Standard polymerase chain reaction (PCR) amplifications were performed with the primer pair: LCO1490 + HCO2198 (Folmer et al. 1994) or C1-J-1718 (Simon et al. 1994) + HCO2198 (for the Neotropical material, see Table 1). The PCR products were purified using the QIAquick PCR purification kit (Qiagen) or with ExoSAP-IT PCR Product Cleanup (Affymetrix, Ohio, USA), and customary sequenced from both directions by Macrogen (Amsterdam, The Netherlands) or Microsynth (Balgach, Switzerland). Forward and reverse sequences were assembled and edited using Geneious R7 v.7.1.3 (Biomatters Ltd).

Analyses of molecular data. To get an overview of the phylogenetic position of the here newly described species and its closely related species, we reconstructed a Bayesian phylogenetic tree. The molecular analysis comprised sequences of 70 specimens, including 21 newly sequenced specimens (Table 1). A multiple sequence alignment was obtained by using MAFFT v.7.050b (L-INS-i algorithm with default settings; Katoh and Standley 2013). The alignment was checked for the occurrence of stop codons and indels using Mesquite v.2.75 (Maddison and Maddison 2011). Identical haplotypes were removed using collapsetypes_v.4.5.pl (Chesters 2013). An ultrametric gene tree was reconstructed using a coalescent prior sensu Monaghan et al. (2009) and Rutschmann et al. (2014, 2017) in BEAST v.2.4 (Bouckaert et al. 2014). We applied the HKY + Γ model of molecular evolution as this has been shown to fit well cox1 data sets of mayflies (e.g. Rutschmann et al. 2017, Gattolliat et al. 2018). Moreover, more complex partitioning-schemes resulted in very low effective sample sizes (ESS) for preliminary tree inferences (see also Gattolliat et al. 2018). Three independent tree inferences each with 50 mio generations were performed in order to check consistency among runs. The convergence of each run and between runs was checked using Tracer v.1.7 (Rambaut et al. 2018), considering effective sample sizes > 200. We summarized the trees using LogCombiner v.2.4 (Bouckaert et al. 2014), discarding 10 % of each tree as burnin and obtained a maximum clade credibility tree using TreeAnnotator v.2.4 (Bouckaert et al. 2014).

Recently, Salles et al. (2016) performed a comprehensive phylogenetic analysis of Cloeodes and related taxa. This reconstruction had considerable impacts on the systematics of the group: Cloeodes is restricted to Neotropical and Nearctic regions; all the African species were re-assigned to Potamocloeon or Crassolus; Oriental species were assigned to one of the three subgenerea of Bungona. Based on its morphology, the new species is attributable to the Cloeodes-complex sensu Salles et al. (2016), sharing all diagnostic characters (apex of maxillary palp slightly narrowing between one-half and apical one-third, presence of a row of long conspicuous setae at the base of tibiae and on tarsi, presence of a row of conspicuous long setae on the abdominal sterna, and apex of gill III or IV pointed).

Within the Cloeodes-complex, B. (Ch.) pontica sp. n. shows all the features delimiting the genus Bungona as defined by Salles et al. (2016): antenna short (Fig. 1B); spine-like setae between prostheca and mola of right mandible present (Fig. 2C); labial palp segment III distally expanded (Fig. 2E, G); setae on the margin of femur long (Fig. 3A, C, E); patella-tibial suture absent on fore tibia and located at middle length of the middle- and hind tibia (Fig. 3B, D, F); a row of conspicuous setae on the dorsal surface of the fore tibia exceeding half of the length of tibia (Fig. 3B); a row of conspicuous long setae present on abdominal sterna IV or V–VI (Fig. 5A); distance between the insertions of conspicuous long setae on the abdominal sterna greater than diameter of insertions (Fig. 5A); and secondary swimming setae on the outer margin of cerci present (Fig. 1A).

Bungona (Ch.) pontica also exhibits a characteristic shape of the claw with two rows of flattened denticles (Fig. 4E), a contrasting body colour pattern (Fig. 1A), and lanceolate scales on the surface of the body (Fig. 4A); all these characters are typical for the subgenus Chopralla (Salles et al. 2016). In fact, the only synapomorphy of Chopralla which is absent in the new species is the bifid shape of the right prostheca (Salles et al. 2016: fig. 4C–E). In B. (Ch.) pontica sp. n., as in B. (Ch.) liebenauae (Soldán, Braasch & Muu, 1987) from Vietnam, the slender medial process of the right prostheca is absent (Fig. 2C) and might represent a synapomorphy of this possibly closely related species or might be the result of independent reduction in both species. Despite their similarities, B. (Ch.) pontica can be separated from B. (Ch.) liebenauae by having shorter setae on the dorsal margin of the femur (their length reaching two-thirds of the femur width in B. (Ch.) liebenauae and one-quarter of the femur width in B. (Ch.) pontica) and by the absence of characteristic tubercles on the pronotum (present in B. (Ch.) liebenauae; Fig. 6).

According to the molecular reconstruction, it is clear that B. (Ch.) pontica sp. n. represents an isolated lineage, not forming a well-supported clade with any of the taxa included in the analysis (Fig. 9). Due to the general lack of support especially in the deeper nodes of the cladogram, the phylogenetic position of B. (Ch.) pontica sp. n. within the Cloeodes-complex remains unsolved. The presumed position based on the morphological data is near the clade containing Bungona (Chopralla) sp. from Sumatra and B. (Ch.) ceylonensis from India. These two species formed a well-supported monophyletic group together with D. improvida from Madagascar (Fig. 9). This study represents the first analysis of molecular data within the Cloeodes-complex. A fine-scaled geographic taxon sampling combined with the use of additional markers will be needed to resolve the phylogenetic position of B. (C.) pontica sp. n. and more in general to solve the phylogenetic relationships within the Cloeodes-complex.

Bungona (Chopralla) pontica sp. n. represents a tropical mayfly lineage in the Asia Minor fauna. Within the Cloeodes-complex, B. (Ch.) pontica sp. n. exhibits morphologically high similarity to the Oriental genus Bungona, which is distributed in Australia, New Guinea, and southern, eastern and southeastern Asia (Marle et al. 2016, Salles et al. 2016). The closest morphological relative is B. (Ch.) liebenauae, described from Vietnam (Soldán et al. 1987). Although not supported by our molecular approach, an Oriental origin of B. (Ch.) pontica sp. n. seems most likely, based on morphological data.

The discovery of this new species in Asia Minor is rather unexpected, as other Turkish mayfly taxa have Palaearctic or West Palaearctic distributions (Salur et al. 2016). Only a few taxa with tropical affinities have been reported from Turkey and the Middle East until now; however, most of them should be considered as misidentifications or at least as questionable observations (see below).

The heptageniid Thalerosphyrus determinatus (Walker, 1853) was recorded from Turkey based on findings of Demoulin (1965) and Berker (1981), although the genus Thalerosphyrus Eaton, 1881 is considered endemic to the Oriental realm (Webb and McCafferty 2008, Sartori 2014). Demoulin (1965) noted the resemblance of two immature larvae from near Ankara with T. determinatus as redescribed by Ulmer (1939) from Southeast Asia. Berker (1981) reported T. determinatus from the vicinity of the town of Elazığ but based this on the superficial similarity with Demoulin’s (1965) figures. These figures (Demoulin 1965: fig. 4) do not show any of the diagnostic characters of Thalerosphyrus as defined in Webb and McCafferty (2008), and the specimens might represent some species of Electrogena or Heptagenia. Thalerosphyrus determinatus was included in recent Turkish checklists (Kazancı 2001, Kazancı and Türkmen 2012, Salur et al. 2016) but were highlighted as doubtfully occurring in Turkey. We conclude the occurrence of T. determinatus in Turkey is most probably based on misidentifications.

Another tropical mayfly lineage with a distribution pattern extended into the Middle East is the genus Clypeocaenis Soldán, 1978 (Caenidae), with most of its diversity occurring in the Oriental (Vietnam, Borneo, Sri Lanka, and India) and Afrotropical (Burkina Faso and South Africa) realms (Malzacher 2013). Clypeocaenis bisetosa Soldán, 1978 was partially collected in the Middle East (part of the type series comes from northern Iran, the other from India). Although the species was not recorded in Turkey, it undoubtedly represents an Oriental tropical element in the Middle East mayfly fauna (Soldán 1978).

In the family Baetidae, the fauna of the Arabian Peninsula shows clear affinities with the Afrotropical fauna. Cloeon smaeleni Lestage, 1924, is widely distributed in the Afrotropics; its distribution covers practically all of sub-Saharan Africa, including Madagascar and Reunion Island (Gattolliat and Rabeantoandro 2002), and it was recently reported as invasive in South America (Salles et al. 2014). In the Middle East, it was also reported from Yemen (Gillies 1985) and Saudi Arabia (Salles et al. 2014), but these populations probably represent a separate closely related species (Salles et al. 2014). Recent integrative studies have proven that the Afrotropical Labiobaetis glaucus (Agnew, 1961) also occurs along the south-western coast of Saudi Arabia (Gattolliat et al. 2018). However, these Afrotropical lineages reach their northern limit in the southern Arabian Peninsula and appear absent in Turkey and the Levant. Baetis (Rhodobaetis) pseudogemellus Soldán, 1977, was described from north-eastern Sudan (Soldán 1977) and was subsequently reported from south-eastern Turkey by Kazancı (2009) based on the observation of three larvae. Nevertheless, the identification of this material as B. (R.) pseudogemellus is highly questionable, taking into account that numerous similar species of Rhodobaetis occur in Turkey and surrounding areas (Godunko et al. 2015, Yanai et al. 2018).

Components of the Turkish fauna with tropical affinities were already described also in other freshwater invertebrates, including anostracans, caddisflies, sponges (Banarescu 1990, 1992), and most probably also gastropods (Heller 2007). Furthermore, many species of freshwater fish from the Middle East show an Oriental origin (Heller 2001), as well as several lineages of terrestrial invertebrates (Yom-Tov and Tcherov 1988). The tropical Oriental taxa were supposed to have invaded the Middle East westwards from the Euphrates–Tigris basin (Heller 2007).

Bungona (Ch.) pontica n. sp. is the first mayfly species showing this unusual biogeographic pattern. The species is probably very rare in Turkey; our results only include six individuals from two neighboring localities, despite our extensive collecting throughout the country. It was also not found by other Turkish researchers (Salur et al. 2016) and has not been recorded in neighboring Georgia (Gabelashvili et al. 2018), Iran (Bojková et al. 2018), or Syria (Koch 1988). It might represent a relict population left after previous connection to the Oriental region.