The study was conducted in the central region of the East European Plain, mainly around Moscow, which includes a wide range of habitats from natural forests to urban lawns. The territory is located in a belt of mixed and broad-leaved forests. The climate is temperate continental, the average annual amplitude of temperature variation is 28 °C, and 600–800 mm of precipitation falls per year (Kulbachevskii 2021). Three types of habitats representing different disturbance were selected: natural forests (10 locations), forest parks (six locations), urban lawns (five locations). One 2-litre sample of the litter and / or topsoil was taken from 1 × 1 m area in each location. Samples were taken from 21 locations from October 2020 to February 2021 (Appendix 1). Springtails were extracted into 96% alcohol using Tullgren funnels, without heating, until the samples were completely dry (7-10 days).

The extracted material was sorted under a stereomicroscope. The possible mixing with coexisting congeners of P. notabilis was considered. Apart from P. notabilis, four species of the genus Parisotoma have been recorded in the East European Plain: P. agrelli Delamare Deboutteville, 1950, P. ekmani Fjellberg, 1977, P. reducta Rusek, 1984, and P. trichaetosa Martynova, 1977; all four species are rare and only the latter species was recorded in the Moscow region previously. In appearance all four species are easy to separate from P. notabilis by having an almost white corpus and smaller eye spots. The first three congeners occur only in northern areas and are rarely recorded in the central region of the East European Plain. The littoral P. agrelli lives only on the Arctic Ocean shore, while the Asiatic P. reducta is distributed in the very north-east corner the East European Plain. The boreal P. ekmani can probably occur towards the central part of the East European Plain via peat-bogs although it was never recorded there in spite of intensive study of the region (its distribution is given in more detail in Potapov 1991, 2001). Parisotoma trichaetosa was formally recorded in the Moscow region (Potapov et al. 2021). It is an invasive Asiatic species (Potapov and Janion-Scheepers 2019) with two known single records in the area under study. Parisotoma trichaetosa is distinct by its overall appearance, quadridentate mucro, and many other characters. The differentiating characters of P. notabilis and its congeners mentioned above are given in Table 1. Apart from P. notabilis, other species of Parisotoma were not recorded in our samples. Our preliminary sorting of specimens of P. notabilis was confirmed by the identification by the ribosomal 28S gene region. Besides, the specimens from several samples of each lineage were mounted on slides and identified using existing identification keys (Potapov 2001; Fjellberg 2007).

The lineages of this species can be identified by both the mitochondrial COI gene and the ribosomal 28S gene (D2 region, Porco et al. 2012a; D3–D5 region, Saltzwedel et al. 2017). We used only the D3–D5 region of 28S which showed higher positive PCR products. Ninety-seven individuals from 21 locations were selected for analysis. DNA extraction was performed using the Thermo Scientific Phire Tissue Direct PCR Master Mix. DNA was extracted from single individuals in 20 μl DNA Dilution Buffer and 0.5 μl DNA Release Additive and incubated at 98 °C for 2 min. This technique allows DNA extraction with relatively little damage to the original material, enabling vouchers to be deposited. A 573 bp D3–D5 fragment of the nuclear 28S rDNA was amplified using the primers 28Sa 5’-GAC CCG TCT TGA AGC ACG-3’ and 28Sbout 5’-CCC ACA GCG CCA GTT CTG CTT ACC-3’ (Whiting 2002; Prendini et al. 2005). COI was amplified using the primers ColFol-for 5’-TTT CAA CAA ATC ATA ARG AYA TYG G-3’ and ColFol-rev 5’-TAA ACT TCN GGR TGN CCA AAA AAT CA-3’ (Ramirez-Gonzalez et al. 2013). This data was used to correlate the results with L4-Hebert, for which there are no data on D3–D5 fragment of 28S. To a master mix consisting of 7 µl of nuclease-free Water, 10 µl of Phire Tissue Direct PCR Master Mix, 1 µl of forward and reverse primers were added 1 µl of individual DNA. PCR conditions included one initial activation step at 98 °C for 5 min, followed by 30 amplification cycles of denaturation at 98 °C for 5 s, annealing at 57 °C (28S) or 55 °C (COI) for 5 s, elongation at 72 °C for 20 s and a final elongation step at 72 °C for 1 min. The result of the PCR was evaluated by electrophoresis in agarose gel with ethidium bromide. PCR products were purified using an enzyme mix of 0.5 µl of exonuclease I (Exo I) and 1 µl of recombinant alkaline phosphatase (rSAP) per 5 µl of PCR product, according to the protocol: 37 °C for 15 min and 80 °C for 15 min. The purified product was dried with the addition of the forward primer. Sequencing was performed in the Synthol laboratory. The sequences obtained were edited in Chromas Lite (v. 2.6.6) (http://technelysium.com.au/wp/chromas/). Sequences were then aligned using BioEdit (v. 7.2) (Hall et al. 2011). Tree construction and calculation of genetic distances between lineages (K2P-pairwise distances) were performed using the MEGA-X program (v.11) (Tamura et al. 2021). Tree calculation was performed with the Maximum Likelihood method with the Jukes-Cantor + Gamma Distributed parametric model proposed by MEGA-X (lowest BIC = 4531.333, AICc = 2599.369) for the 28S gene, and the Neighbor-Joining method with Tamura 3-parameter + Gamma Distributed parametric model for the COI gene. Sequences were obtained for 87 individuals for a D3–D5 fragment of the 28S gene 573 bp and 12 individuals for a COI gene 619–657 bp, which are additional confirmation of the results. The tree for the 28S gene is based on our data and those of Saltzwedel et al. (2017). GenBank (www.ncbi.nlm.nih.gov/GenBank) accession numbers are listed in the Appendix 1. Detail data on genetic lineages records were placed in the international Global Biodiversity Information Facility (GBIF) in the ‘sampling event dataset’ format (Striuchkova et al. 2022).

The constructed tree revealed three genetic lineages of P. notabilis: L1, L2, and L4-Hebert (Fig. 1). The last one is not represented for D3–D5 region of 28S in GenBank. It was identified by COI according to blast results in the BOLD database (Hebert et al. unpub. data, Anslan and Tedersoo 2015). Genetic distances (observed K2P-distances in %) between the lineage L4-Hebert (this study) and L4-Saltzwedel for COI gene = 17.6%. The genetic distances between the L1, L2 and L4-Hebert lineages for the D3–D5 region of 28S ranged from 0.9% to 1.4% in our material. Mean genetic distances between all known lineages and intralineage distances are given in Table 2. Neighbor-Joining genetic tree and intralineage and interlineage K2P-pairwise distances for the COI gene are presented in Appendices 2 and 3, respectively.

Figure 1. 

Maximum Likelihood genetic tree of six lineages of P. notabilis based on the D3-D5 region of 28S gene fragment (Bootstrap support values shown on the branches, scale bar shows genetic distance) including data from Saltzwedel et al. (2017) availed in GenBank. Our data is highlighted with colored vertical lines.

Numerous cases of sympatry of the lineages were revealed (Fig. 2). Two lineages were found in one third of the locations (41%), with three lineages in 12% of the locations. Almost half (47%) of the locations contained only one lineage. Table 3 shows the ratios of the number of sites with lineage sympatry / number of total sites sampled based on the literature (Porco et al. 2012a; Saltzwedel et al. 2017) and our data.

Figure 2. 

Records of the genetic lineages of Parisotoma notabilis (green circle - L1, orange square - L2, red triangle – L4-Hebert). See Appendix 1 for location data for each lineage.

L1 and L4-Hebert lineages were recorded in the urban areas, and all three lineages - in the partly disturbed (forest parks) and undisturbed (forests) habitats. In the forests, the occurrence of L2 and L4-Hebert lineages was about the same, while L1 was sporadic. In forest parks, L1 was the most common, the L4-Hebert was less common, while L2 was rare. In the city (urban areas), L1 absolutely dominated (Fig. 3).

Figure 3. 

Records of genetic lineages of P. notabilis (L1, L2 and L4-Hebert) in different groups of habitats, % of the number of samples in the habitats of the group.

The L1 and L2 lineages of P. notabilis were already cited in the East European Plain (Saltzwedel et al. 2017). The lineage L4-Hebert has been identified based on the sequences COI available in GenBank (Hebert et al. unpub. data; Anslan and Tedersoo 2015). The L4-Hebert for D4–D5 region of 28S is most similar with the L0 and L3 lineages (0.50%) which have not yet found in the East European Plain. The L4-Hebert lineage showed the greatest differences with the L1, L2, and L4-Saltzwedel lineages (0.90–1.05%). Updated data in the present study agree with the previously estimated average genetic distance between the most common lineages of P. notabilis L1 and L2 in Europe and Canada (1.4% vs. 1.39% according to Saltzwedel et al. 2017).The L4-Hebert lineage is probably widespread in the East European Plain, because we recorded it also in Voronezh, which is about 500 km south of Moscow.

Previously, studies on the genetic structure of P. notabilis populations showed mainly parapatric distribution of lineages. The simultaneous presence of two or three lineages was noted only in 13% of the total number of locations studied (Porco et al. 2012a; Saltzwedel et al. 2017). Consequently, the sympatry of the lineages in Canada was thought to be possible independent introductions of the species, that is, they consider the case likely to be accidental (Porco et al. 2012a). In our data, the sympatry of the lineages showed markedly more frequency than in the west - almost in half of the locations studied (Table 3). The ratio number of sites with lineages sympatry / number of total sites sampled is greater in our findings. In Western and Central Europe this value is 0.8–0.18, while in the East European Plain it is 0.5. The joint occurrence of lineages was noted not only in different areas of the Moscow region, but also in Voronezh. In Eastern Europe, sympatry of the lineages L2 and L4-Hebert was also found in Estonia (Anslan and Tedersoo 2015).

Presumably, the number of detected lineages in one location may depend on the number of individuals analyzed from a core. Our data shows, however, that even small samples of 4 or 5 individuals sometimes revealed up to three genetic lineages (Fig. 4), while a sample of 17 individuals in Hamburg (Germany) was represented by only one lineage (Porco et al. 2012a). Our material is collected over a relatively small area compared to extensive data from Western Europe. It is possible that the urban condition contributes to the conservation of invasive species (Rebele 1994) and also should concern genetic lineages. Nevertheless, it is hard to ignore that several lineages are found in true natural forests (Shakh-2021-2) located more than 100 km away Moscow city. It is likely that the relatively separated lineages of P. notabilis in Western and Central Europe do co-occur in the east. One of the reasons for this phenomenon may be the geologically young postglacial landscape of the region we studied. This flat area had no refugia and broke free from ice only a few tens of thousands of years ago (Velichko et al. 2004) and so it is still an arena of active migratory flows and evolutionarily young communities, which is especially evident in the sedentary groups of organisms (Markova et al. 2008). The species introduced into such young communities are more successful than those introduced into ancient communities. A similar explanation can supplement the discussion of the reasons for the sympatry of P. notabilis lineages in Canada as a result of their introduction to North America (Porco et al. 2012a; Saltzwedel et al. 2017). Our hypothesis on parapatry of the lineages in the East European Plain has not been confirmed.

Figure 4. 

Dependence of the number of detected P. notabilis lineages on the number of individuals analyzed in a sample in Europe. The larger the circle, the more samples contain a certain combination of the number of lineages and the number of individuals.

At present, it is difficult to conclude how widespread sympatry of genetic lineages is. For example, sympatry of genetic lineages was found in Deutonura monticola (Cassagnau) and Heteromurus major (Moniez) among 16 genetically studied Collembola species (Porco et al. 2012b).

Genetic studies of springtails often do not consider the habitats where the material was collected from, or, at least, do not specify them. Saltzwedel et al. (2017) preliminarily supposed that environmental conditions may select for different lineages of P. notabilis. We investigated the genetic composition of P. notabilis populations not only in natural but also in disturbed habitats. The results showed an association of the L1 lineage with disturbed habitats but L2 and L4-Hebert with natural habitats. This indicates possible differences in tolerance to disturbance among genetic lineages.

The different habitat preferences of genetic lineages reflect the process of ecological diversification within one species, and can lead to the emergence of new species resulting from ecological speciation. For many taxa, such cases were noted based on the action of different selection vectors according to the gradient speciation model (Doebeli and Dieckmann 2003; Endler 2020). This phenomenon was found, for example, in the “palustris” group of the genus Isotomurus, the genetic analysis of which made it possible to distinguish six species within I. palustris (Mueller) sensu lato distributed in different sections of the river floodplain profile (Carapelli et al. 1995). Subsequently, broad genetic variability was discovered in many Collembola species, but the cryptic diversity of this taxon was mostly described in terms of genetic lineages rather than independent species (Porco et al. 2012a). Obviously, more genetic, morphological, and ecological information must be accumulated to standardize the criteria for distinguishing Collembola species.

Information on the ecological preferences of genetic lineages is sporadic. Thus, in the widespread springtail species Lepidocyrtus lanuginosus the genetic lineage L1 was abundant and occurred in each of the three habitats studied (forests, grassland, arable fields), L2 only in forests, and L3 only in pastures and arable fields (Zhang et al. 2018). A similar diversification is found in P. notabilis: one lineage is unspecialized to habitats (L4-Hebert), while two are more specialized, one to natural forests (L2), the other to disturbed habitats (L1). In both species the environmental factors leading to lineage specialization remain unclear. The adaptations of the lineage may be of ecophysiological, nutritional, demographic or migrational nature.

Our results are preliminary and call for more data. However, the assumption that different genetic lineages of P. notabilis prefer different degrees of habitat disturbance seems convincing at this point. Samples along the disturbance gradient can reveal the diversity of lineages for one location. The cosmopolitan species P. notabilis, abundant in natural and disturbed habitats, is a promising model object for studying the phylogeography of Collembola populations.