Adult mayflies were collected from five distinct localities of Hokkaido in 2019 and 2020 (Fig. 1). All of them were collected by using a light trap or collected around street lights. They were stored in 80% ethanol. Genital structures were observed under a Zeiss Axiophoto microscope (CarlZeissAG, Jena, Germany). Digital images of specimens were captured using AxioCam ERc 5s (CarlZeissAG, Jena, Germany) attached to the microscope. The genital structures were observed after processing in Proteinase K and stained with acid fuchsin. The mesonotum structure was observed after processing in potassium hydroxide (KOH). Eggs were extracted from a female imago from Loc. 2 and dehydrated through a series of increasing concentrations of ethanol (80%, 90%, 95%) before immersion in absolute ethanol, 10 minutes each, then placed in hexamethyldisilizane (HMDS) for 15 minutes. After 30 minutes air drying, eggs were mounted on stubs with conductive sticky tape and sputter-coated with gold-palladium in an ion-sputtering device MSP-20-MT Automatic Magnetron Sputter (Vacuum Device, Ibaraki, Japan). Finally, the eggs were observed through scanning electron microscopes JSM-5310LV (JEOL, Tokyo, Japan). The terms of morphology followed Bauernfeind and Soldan (2012) and the terms of genital morphology followed Malzacher (1991).

Figure 1.  

Locations of collecting localities. Loc. 1 Usotan, Hamatonbetsu-cho, Hokkaido, Japan. N45.046779, E142.361674; Loc. 2 Kita 17 jo, Nishi 9 chome, Kita-ku, Sapporo-shi, Hokkaido, Japan. N43.079981, E141.338084; Loc. 3 Rankoshi, Chitose-shi, Hokkaido, Japan. N42.809656, E141.574878; Loc. 4 Poropinai, Chitose-shi, Hokkaido, Japan. N42.801191, E141.326650; Loc. 5 Shirooka, Assabu-cho, Hokkaido, Japan. N41.832767, E140.332714.

Total DNA was extracted from the specimens using Qiagen DNeasy Tissue Kit (QIAGEN, Venlo, The Netherlands). DNA of male adults was extracted from the whole abdomen. The exoskeleton was kept for further observation. DNA of female adults was extracted from the thoracic muscle tissues. The barcode region of COI gene was amplified by the PCR method using the primer set, LCO1490 and HCO2198 (Folmer et al. 1994). The PCR cycle was: 1 min at 94°C followed by 40 cycles of 30 sec at 94°C, 30 sec at 45°C, 1 min at 68°C, then 1 min at 68°C using BIO RAD T-100 Thermal Cycler (Bio-Rad Laboratories, California, USA). PCR products were purified using Fast Gene Gel/PCR Extraction kit (Nippon Genetics, Tokyo, Japan). The purified DNA was sequenced using Beckman Coulter CEQ2000XL (Beckman Coulter, California, USA). The COI sequences of three male adults and five female adults from Loc. 1, 2 and 5 were deposited to GenBank under accession numbers as noted in Materials, Taxon treatment.

All sequence data were aligned automatically using MAFFT v.7.429 (Katoh et al. 2002) then visually cross-checked using Mesquite software (Maddison and Maddison 2019). Pairwise distances were calculated using MEGAX (Kumar et al. 2018, Stecher et al. 2020). A substitution model was selected using ModelFinder (Kalyaanamoorthy et al. 2017). Maximum Likelihood analysis was conducted using IQ-Tree (Nguyen et al. 2014). Ultra-fast Bootstrap was conducted (Hoang et al. 2018) with 1000 replications. The Maximum Likelihood tree was edited in the FigTree v. 1.4.4 (http://tree.bio.ed.ac.uk/software/figtree). The sequences of the related taxa were obtained from an article proposing DNA barcode library, based on taxonomically well-curated specimens (Morinière et al. 2017).

The genital characters of the Japanese specimens agree well with those of the European specimens as described in Malzacher (1986) and Malzacher (1984): the caudal part of styliger roundly convex, forceps strongly bent in middle and strongly sclerotised in tip, penis anvil-shaped and widens in apex. The shape of penis lobe varies as straight rear edge to V-shaped (Malzacher 1984, Malzacher 1986), but this variation depends on the contraction of the penis-muscles (Malzacher 1991). Our specimens have straight or slightly bent rear edge as in Fig. 4a, b and c. Malzacher (1986) described the lateral-sclerite, the base of forceps and the styliger-sclerite rarely tinted brownish and our Japanse specimens show this colouration (Fig. 4b). All other articles describing the genital characters of this species also show long forceps bending inwardly and penis lobe widens in apex (Saaristo 1966, Zhou and Zheng 2004, Zhou et al. 1997, Bauernfeind and Soldan 2012) and do not contradict the Japanese specimens.

According to the morphological observations, we identified the Japanese samples as C. rivulorum, as is noted in Diagnosis, Taxon treatments. However, the genetic difference between Japanese (Hokkaido) samples and European (German) C. rivulorum is significantly larger (> 11%) than 2.2%, which has been generally recognised as a level of divergence delimiting species across diverse insect taxa (Hebert et al. 2003). The genetic difference may indicate a possibility that they should be designated as independent species. Nevertheless, we could not find any morphological differences between the Japanese samples and the European population of C. rivulorum, as described in Malzacher (1982), Malzacher (1984), Malzacher (1986). Large intraspecific variations of COI gene were also identified in Cloeon dipterum Linnaeus, 1761, the widely distributed baetid species known to be a habitat-generalist (Rutschmann et al. 2014, Rutschmann et al. 2017, Yano et al. 2019), in which seven distinct clades showing a maximum of ca. 13% of uncorrected genetic differences were identified. Amongst them, four Eurasian clades (named "CT1", "CT2", "CT3" and "JK") show at most 9.1% ("CT3", northern Europe vs. "JK", Japan and Korea) p-distance. Caenis rivulorum also inhabits a wide range of habitats (Bauernfeind and Soldan 2012) and is distributed from the east to the west part of Eurasia so that, as in the case of Cloeon dipterum, the genetic distance can be high when compared with distantly-located populations. In the case of Cl. dipterum, the larger genetic distance between Eurasian clades and Macaronesian clades (named "IS1", "IS2" and "IS3") have been shown (Yano et al. 2019) and those Macaronesian populations regarded as being distinct species (Rutschmann et al. 2014, Rutschmann et al. 2017). Future detailed studies might reveal the structures of populations and speciations; however, we currently do not have any reason to separate morphologically indistinct populations as independent species, which may even cause taxonomic confusions.

The Japanese fauna of the genus Caenis are poorly studied. The following two named species have been reported from Japan to date (Ishiwata 2018):

1. Caenis horaria (Linnaeus, 1758) from Honshu. This species was reported by Ueno and Okamoto (1932) in Nippon Kontyu Zukan ("Iconographia Insectrum Japonicorm"). However, only a brief description of the species was given in the article. This species is morphologically similar to C. rivulorum. Therefore, we need further consideration for the record of this species from Japan.
2. Caenis nishinoae Malzacher, 1996 from Lake Biwa. This species was described by Malzacher (1996), in which detailed descriptions of all stages of both sexes are given. This species can be clearly distinguished from C. rivulorum.

In addition, a couple of unidentified species have been reported, based on larval stages: Caenis sp. CA and Caenis sp. CB (Gose 1958). We can assume that those species are not C. rivulorum for the following reasons: the lateral margin of pronotum in Caenis sp. CA expands anteriorly and forms a tip apically (Gose 1958), but is straight or slightly convex and more or less parallel in C. rivulorum (Bauernfeind and Soldan 2012). Caenis sp. CB does not have short strong bristles on its fore-femora (Gose 1958) in contrast to 3–7 relatively short strong bristles on fore-femora in C. rivulorum (Bauernfeind and Soldan 2012).

Ishiwata (2001) reported Caenis sp. with a description of the adult structures, but the genital structures were not described nor illustrated. According to the drawing in Ishiwata (2001), the antennal and sternum structures are similar to those of C. rivulorum examined here. However, we cannot determine that Caenis sp. is C. rivulorum, based only on the published information.

Ishiwata and Fujitani (2018) described two SEM-image of eggs of unidentified species of Caenis: Caenis sp. 1 and Caenis sp. 2. The micropyle of those eggs is straight to aperture which is in contrast to that of C. rivulorum.

Recently, an additional unidentified species of the genus was reported from Ura-Bandai, Fukushima, Japan (Masubuchi and Tsutsumi 2015, Masubuchi and Tsutsumi 2014), which shows a unique shape of the male genitalia. Detailed morphological and ecological information have been provided for this species, but a taxonomic name has not yet been proposed. This species can clearly be differentiated from C. rivulorum, based on the male genitalia.