Basic description and some notes on the evolution of seven sympatric morphs of Dolly Varden Salvelinus malma from the Lake Kronotskoe Basin

Abstract The study examines the basic morphological and ecological features of Dolly Varden from Lake Kronotskoe (Russia, Kamchatka). Seven valid morphs different in head proportions, feeding, timing, and place of spawning have been determined in this ecosystem. The basic morphometric characteristics clearly separate Lake Kronotskoe morphs from each other, as well as from its potential ancestor (Dolly Varden). According to CVA analysis, the most notable morphological characteristics determining the mouth position are the length of a lower jaw and rostrum. Furthermore, five of seven morphs inhabit different depth zones of the lake and feed on different food resources. Our data suggest that reproductive isolation may be maintained by temporal/spatial isolation for two morphs with lacustrine spawning, and by spatial isolation only for the rest of the morphs with riverine spawning. The sympatric diversity of the Lake Kronotskoe charrs is exceptionally wide, and there are no other examples for seven sympatric morphs of genus Salvelinus to coexist within a single ecosystem. This study puts forward a three‐step hypothetical model of charr divergence in Lake Kronotskoe as a potential ground for future studies.

The phenotypic variability of northern Dolly Varden is mainly realized through various strategies ranging from anadromous to stream resident (Mochnacz, Schroeder, Sawatzky, & Reist, 2010;Pichugin, 2015). The lakes are typically used for growing, yet not for spawning.
However, due to volcanic activity in the North Pacific region, numerous lakes were formed by lava dams. Some of them are inhabited by landlocked Dolly Varden that adapted to the lacustrine conditions. Dolly Varden in these lakes is typically represented either by omnivorous slow-growing populations or exhibits horizontal transformations sensu Adams (1999) into "normal" morph and cannibals. The only complex variant of sympatric divergence in Dolly Varden was found in Lake Kronotskoe to the East of Kamchatka. Different studies report on 3 to 7 morphs in this ecosystem (Markevich, Esin, Saltykova, et al., 2017;Pavlov, Kuzishchin, Gruzdeva, Senchukova, & Pivovarov, 2013;Viktorovsky, 1978). Moreover, they were proved to originate from anadromous Dolly Varden through genetic analyses (Ostberg et al., 2009;Senchukova, Pavlov, Mel'nikova, & Mugue, 2012). Seven distinct morphological types have never been observed previously for any sympatric complex in genus Salvelinus that makes the investigation of morphological and ecosystem features of the Lake Kronotskoe morphs necessary for understanding more complicated mechanisms of diversification.
Meanwhile, little is known about the morphology of these morphs as well as their feeding and spawning biology. As it was mentioned above, Dolly Varden is well-adapted to anadromy and usually spawns in rivers or brooks. Conserving these features in case of isolation in the Lake Kronotskoe basin may have resulted in specific diversification patterns which were not previously observed both in Arctic charr and Lake trout.
Consequently, the objective for this study was to revise the diversity within the charr morphs from the Lake Kronotskoe basin based on their ecological and morphological features. Outlining the mechanisms of diversification in Lake Kronotskoe may expand the understanding of the "charr problem" in general.

| Study area
Lake Kronotskoe is situated on the eastern shore of the Kamchatka peninsula in the Kronotsky Nature Reserve. The basin area is 246 km 2 , the average (max) depth is 58 (136) m, the a.s.l. is 372 m, the catchment area covers approximately 2,330 km 2 . The lake has a developed tributary system comprising both small plain-type brooks with the altitude difference from the head to the mouth approximating to 30-50 m as well as big branched rivers with the drop height of approximately 300-500 m. The ecosystem originated 12-14 thousand years ago due to the volcanic eruption when lava had blocked the valley of the ancestral river thus forming the lake in the upper part of the basin (Braitseva, Melekestsev, Ponomareva, & Sulerzhitsky, 1995).
The Kronotskaya River flows down from the lake through the rapids impassable for anadromous fishes (Kurenkov, 1977;Viktorovsky, 1978).
The lake belongs to the dimictic type: the process of water mass mixing occurs in June and at the end of October; the epilimnion reaches 20-25 m at the end of August. Around 80% of zooplankton biomass concentrates in this upper zone. The shores, littoral zone, and small islands are formed by lava boulders from the surrounding volcanoes. The shallow stony areas are characterized by high production of gammaruses, snails, chironomidae, and stone-fly larvae. The profundal zone is covered with soft grounds; clams, oligochaetes, and chironomidae larvae abound (Kurenkov, 1978).
The fish community is represented by two species: Dolly Varden and kokanee (landlocked Oncorhynchus nerka [Walbaum, 1972]). Both species exhibit the trophic-based diversification and jointly use all the food resources available in the lake (Kurenkov, 1977;Viktorovsky, 1978). All charr morphs were traditionally named with respect to the head shape with the sole exception. The first data about predatory longhead (L), benthivorous nosed (N), and omnivorous white (W) morphs were obtained in 1970s (Viktorovsky, 1978). Lately two deepwater morphs-bigmouth (B) and smallmouth (S) charrs were described from the lake basin (Markevich, Esin, Saltykova, et al., 2017).
Two ephemeris morphs (riverine relic Dolly Varden and Dwarf charr) were also described in the ecosystem (Ostberg et al., 2009;Pavlov et al., 2013). In this study, the first one was reduced to an immature W-morph (discussed below), while the second one was described on the four specimens only and have never been caught since; therefore, it is doubtful that it may be associated with any of the seven morphs.

| Field sampling
The material was collected during the ice-free period lasting from June to October 2013 and 2014 ubiquitously across the lake basin. Fish was caught in the 60-90-m lines of multipanel gill nets 20-50 mm mesh size at the depth from 2 to 100 m. The estimation of fish abundance was based on the results of fishing in all relevant habitats: the lower reaches of the tributaries, littoral (depth < 10 m), sublittoral (10-20 m), epilimnion (<20 m), hypolimnion (>20 m), and profundal (>40 m) zones. No less than 20 gill net samplings were performed per zone in various parts of the lake including different bays, areas around islands, banks, and river mouth vicinities ( Figure S1). To get a comprehensive picture, we tried to invest the same fishing efforts in different parts of the lake according to the depth profiles, winds, temperature conditions, etc. (see Figure S1). The fishing operations were performed simultaneously in all zones once in 2 weeks, that is, gill nets were exposed in littoral+sublittoral, epilimnion, hypolimnion, and profundal zones roughly at the same time oat each station according to the scheme represented in Figure S1 for the sites near the shoreline and the banks in the open waters. The exposure time varied as following: in littoral and sublittoral zone it lasted for 3-6 hr, in epilimnion and hypolimnion-12-15 hr, in profundal-6-8 hr. Totally 1,232 specimens were caught. Additionally, 27 prespawners of anadromous Dolly Varden from the nearest river basin (the Komarova River) were sampled in September 2013. Sampling in the Komarova River was carried out by gill nets with 20-55 mm mesh size. All fishes were caught in the lower course of the river before they reached the reproduction sites.

| Primary separation of morphs
The diversity of morphs was initially estimated during the catch checking. All charrs were separated into seven groups according to their head length, mouth position, and coloration (Table 1, Figure 1). N-morphs were further additionally differentiated by the shape of the head. The first group (N1) had the oval shape of the snout, while N2-morph had a sharped rostrum and N3-morph-hypertrophied flat rostrum. Furthermore, N2-and N3-morphs unlike N1-morph were characterized by an uncovered palate (shortened lower jaw), both jaws were covered by connective tissue ("lips"). N3-morph featured the well-developed connective tissue plates that covered the snout and formed a sharpened spade-like rostrum. In N2-morph, the connective tissue plates turned out to be absent. No difficulties were experienced during the primary separation of morphs. All specimens were sorted by groups unambiguously, and no morphological continuum of variation was revealed.

| Morphology analysis
Geometric morphometric methods were applied to analyze the morphological differences between the groups of charrs. Totally, 322 fish samples were randomly subsampled from the Lake Kronotskoe catches (approximately 20% for abundant morphs and 50% for rare ones). All 27 fishes from the Komarova River, which is southward to the Kronotskaya River basin, were analyzed. The left side of these specimens was photographed by Canon EOS 350D with 50 mm lens in the orthogonal projection, fins straightened, and mouth closed.
In addition, the bottom view of the head was photographed for Nmorphs. To circumvent any optical distortions, the longer side of the image was always set to fit at least twice the fish length, and the same values of the focal length (80 mm) were used in imaging. Totally 23 landmarks were digitized for each fish ( Figure S2) using a tpsUtil and tpsDig2 v2.16 software platforms (Rohlf, 2010).
The comparison was performed in three steps. First, the differences among the Lake Kronotskoe morphs were normalized against the anadromous Dolly Varden. Then the differences among five main morphs (all nosed charrs combined), and among three N-morphs were evaluated. The analysis was carried out by MorphoJ 1.06d (Klingenberg, 2008). A generalized Procrustes superimposition (Dryden & Mardia, 1998) was applied to minimize the distances between individual landmark configurations (Zelditch, Swiderski, Sheets, & Fink, 2004). Therefore, the Procrustes distances became relative measurements of the shape differences among the groups, and a general Procrustes dispersion analysis was used to estimate the divergence among the groups (Rohlf & Slice, 1990). Sexual dimorphism was not profound in the measured traits (Procrustes ANOVA F 318,13558 = 12.7, p = .083);

T A B L E 1 Differentiation of morphs in mixed gillnet catches
thereafter, all the calculations were performed on both sexes combined. To cross-validate our preliminary classification Jack-knife was used for an unbiased estimation of the morphs' classification.
Canonical variates analysis (CVA) was applied to assess the total amount of variation in body shape among the groups. This method reduced the amount of variation within the groups and made the variation among them more vivid (Klingenberg, Barluenga, & Meyer, 2003;Zelditch et al., 2004). Multi-dimensional plots were used both to describe the morphospace (i.e., the abstract space where each point represents a particular individual) and to identify the differences between the groups. To evaluate the contribution of dispersion of individual landmarks to the General discrimination of the groups' body shape, canonical loadings of landmarks on the first three axes were estimated. To visualize the shape changes in the groups, the average head and body shapes for the sets of individuals were compared with the consensus shape of all fish subjected to analysis.

| Fish size differentiation
Fork length (FL, mm) and weight (W, g) were measured for all specimens, and multivariate dissimilarities in fish size were statistically examined using post hoc ANOVA Tukey's HSD. The mature status was determined according to Eenennaam and Doroshov (1998) recommendations, wherein all specimens with gonad development over III stage were considered as mature (1), fishes with the gonad development lower than III stage-as immature (0). Then, each set of fish was sorted by the fork length, and the mature statuses were plotted as a function of length. The mature status plots were averaged using 10-point running average procedure, and approximated by a simple exponential function: where M is the average mature status, FL 0 is the minimal length of maturation, and FL e is the characteristic length of maturation. Thus, the average 50% maturation fork length was calculated as:

| Analysis of fish biology
The analysis of stomach content was carried out in the field laboratory within 3-5 hr after the fish sampling. Food objects were identified according to Manko (2016) recommendations by means of stereomicroscope MBS-9 (12-20x). Prey from the stomachs of all the Lake Kronotskoe charrs was sorted into the following groups: chironomidae larvae, chironomidae pupae, stone-fly larvae, insect imagoes, gammaridae (Gammarus spp.), oligochaetes, snails (Anisus spp. and Limnea spp.), clams (Pisidium sp.), fish. The occurrence of these food objects was estimated for charr groups as where N i is the fish number for each morph with the specified prey in the stomach, N tot is total fish number for each morph with the stomach content (Amundsen, Gabler, & Staldvik, 1996).
The observation of the spawning grounds was carried out within the period from August to October of 2011-2015 in the basins of main tributaries (4 rivers and 5 brooks) and till December in the lake.
Additionally, the lake area was studied in March-April 2011 to define the B-morph spawning period.
F I G U R E 1 Sympatric morphs of Lake Kronotskoe charrs. W, white; L, longhead; N, nosed: 1-blunt nosed, 2-sharpnosed, 3shovelnosed; S, smallmouth; B, bigmouth The first step of the observation aimed to define the beginning of the spawning migration in different rivers and brooks. Every 2 weeks, the lower courses (1-2 km) of each river/brook were visually inspected, the prespawning fish was defined by the development of spawning coloration, thus the beginning of spawning migration was determined. The second step focused on determining the reproduction sites and confirming the stability of spawning distribution. All rivers chosen for our analysis were totally inspected via rafting or hiking from the riverheads to the rivermouths no less than 2 times in separate (discrete/different) years ( Figure S1). The third step comprised the regular observations at several spawning sites of all morphs from the beginning to the end of spawning period in order to define the reproduction period and confirm the spawning isolation between the existent morphs ( Figure S1).
Fishing of spawners at all three steps was implemented according to the catch and release principle by fishing rods. Therefore, spawners were not used for any analysis which included fish killing. Spawning sections in the tributaries were determined by visual observations, the numbers of redds occurrence and females on late IV or V maturation stages were presented at the current sites. The identity of morphs was determined by the aforementioned criteria, and all fish were also photographed to get the additional possibility for morph specifying and to be stored in the personal archive. The description of riverine spawning sites included the mean channel gradient calculated as a ratio of the altitude difference to the river section length in m/km. The distribution over the lake area and the depth range was specified for the lacustrine spawning areas. As it was impossible to visually observe the redds being constructed in the profundal zone, the spawning areas as well as the timeframe were determined by the gill net catches. Only fishes sampled on the V maturation stage were counted as spawners.

| Morphological differentiation of morphs
The first step of morphological comparison was to distinguish the specific features in shape for the Lake Kronotskoe charrs as con- F I G U R E 2 Results of charrs morphological comparison derived from CVA, (a) The averaged Dolly Varden morphological shape against Lake Kronotskoe charrs consensus shape; (b) The shape changes of five main Lake Kronotskoe morphs in relation to the consensus body shape; (c) Plot of specimens of five main Lake Kronotskoe morphs in CV1-CV2 morphospace success in Dolly Varden classification against W-morph, only 52% of the anadromous charrs was classified correctly (Table S1).
Herewith, N-morphs had comparatively shorter jaws and a higher nape. S-and B-morphs differed predominantly in eye diameter and jaw length. L-morph was distant from the Dolly Varden in a much lower head, longer snout, gill cover, and jaws. The differences described above can be recognized quite clearly based on linear measurements of charr heads (Table S2).
The next step of comparison indicated that the body shape is different among the five main Lake Kronotskoe morphs (Procrustes ANOVA F 294,1375 = 12.6, p < .001). The most distant groups were deepwater B-morph and epilimnetic L-morph, while W-morph took the closest position to the center of the grid. N-morphs turned out to be closer to W-morph; hypolimnetic S-morph was closer to the second deepwater B-morph ( Table 2). The maximal internal morphological diversity characterized N-morph, whereas L-and B-morphs were minimally polymorphic.
CVA of the landmark positions resulted in outlining seven significant CVs differentiating the morphs (Table S3, Wilk's lambda < 0.001).
The degree of divergence between the morphs is illustrated in CV1-CV2 space (Figure 2c). According to the canonical coefficients (Table   S4) (Table S1) with minimal classification probability for W-morph (84%) and maximal for B-morph (100%).
Clear morphological differences across three N-morphs were also revealed. The use of Procrustes dispersion analysis demonstrated the differentiation among the groups to be evident either in side-view head and body shape (F 88,1620 = 2.01, p < .010), or in ventral-view head shape (F 32,1760 = 11.02, p < .001). Along with it, the relative Procrustes distances between the groups were smaller in comparison with the distances across all five main morphs ( Table 2). The maximal internal morphological diversity characterized N1-morph while N3-morph was minimally polymorphic and more distant from the others.
Variation in head and body shape of N-morphs was associated with three CV roots (Table S3, Wilk's lambda < 0.001). Based on the canonical coefficients of side-view CVA (Table S4) (Table   S1).
Thus, all the groups initially classified by the qualitative characteristics during the gillnetting significantly differed by the head and body morphology. This outcome of our comparison confirms the existence of seven morphs in the ecosystem of Lake Kronotskoe.

| Size and weight differences
Descriptive characteristics of the seven morphs are represented in

| Distribution patterns and feeding
Our investigation revealed that the assigned morphs are characterized by the set of specific features of distribution in the lake ecosystem (Table S5) Adult fish switch to predatory feeding and majorly consume deepwater charrs and kokanee (0.79), while the occurrence of shallow water invertebrates significantly decreases (0.15).

| Spawning
The spawning areas of all seven morphs were documented all over the basin by mapping the reproduction sites either in the tributaries or in the lake. All morphs can be divided into two groups: la-

| DISCUSSION
The complex analysis comprising both the morphological and the ecological characteristics provided the solid ground for outlining seven distinct charr morphs coexisting in the Lake Kronotskoe ecosystem.
The morphological analysis has revealed that six morphs are clearly distant from the anadromous Dolly Varden in head and body morphology; yet, only W-morph has no reliable head shape features which separate it from the anadromous Dolly Varden. It should be noted that the Dolly Varden body shape complied with the consensus while all the other morphs were distant from the consensus in various ways.
Both deepwater morphs, in comparison with epilimnetic ones, had significantly bigger eyes and a shorter snout. The blunt head shape and the large eye are common morphological features of deepwater charrs which were observed for different lakes across the range. In particular, it was previously described for deepwater benthivorous morphs from lakes Skogsfjordvatn (Scandinavia), Kamkanda (Transbaikalia), Loch Dughaill (Scotland) Hooker et al., 2016;Knudsen et al., 2016;Skoglund, Siwertsson, Amundsen, & Knudsen, 2015). However, B-and S-morphs are clearly different from each other in jaw length. The jaws of S-morph are short proportionally to the snout length; thereby, this morph conserves the terminal mouth position, which is common for the omnivorous Dolly Varden. On the contrary, B-morph has a very long lower jaw that once being combined with a short snout results in the supra-terminal mouth orientation.
The most reliable characteristic of the three N-morphs is the subterminal mouth position that is determined by a short lower jaw and a relatively long snout. The shortening of the lower jaw is a trend seldom registered for benthivorous charrs; however, it is common for ciscos and can be regarded as the most wide-spread adaptation for benthic feeding in a littoral zone (Bronte, Fleischer, Maistrenko, & Pronin, 1999;Lu & Bernatchez, 1999;Østbye, Naesje, Bernatchez, Sandlund, & Hindar, 2005). Different N-morphs were clearly distant from each other in the lower jaw length and the head width.
To sum up, all trophic-related morphological adaptations in Dolly Varden morphs from Lake Kronotskoe are formed by the variability in jaw length, which could be either correlated or non-correlated with F I G U R E 4 Spawning sites of charrs in Lake Kronotskoe basin (marked with letters for different morphs) the modifications in the snout length. This variability greatly modifies the head shape and shifts the mouth orientation from terminal to supra-terminal or sub-terminal position thus giving rise to the necessary adaptations to the existing ecological conditions. The only reliable difference in body shape occurs in the length and the height of the caudal peduncles which turned out to be higher in all N-morphs, longer in W-, and S-morphs, and shorter in B-and L-morphs.
Both head and body shape features are strongly associated with the foraging strategy of each morph. The lack of specific adaptations in W-morph compared to the anadromous Dolly Varden corresponds to its broad food niche. Deepwater charrs are defined by clear morphological shifts which are common for all deepwater fishes and were described previously (Alekseyev, 2000;Praebel, Ostbye, Hassve, & Hagenlund, 2015;Skoglund et al., 2015). Therefore, these characteristics can be recognized as a general specialization for foraging in poor in Lake Malawi (Albertson, 2008). Although the similar type of feeding for the three N-charr groups impedes the understanding of the function, the head shape differences are prominent among them. N1morph with the rounded rostrum and well-developed connective tissue on the jaws strongly reminds us of the hypertrophied lips of barbs and cichlids (Colombo et al., 2013;Dimmick, Berendzen, & Golubtsov, 2001;Roberts & Khaironizam, 2009). Such type of head shape was previously recognized as a vital adaptation for effective food extracting from narrow holes between the stones (Lukas, Gonzalo, Frederico, & Axel, 2015). According to the deep structural transformations of skull, we previously assumed that N2-and N3-morphs are specialized groups of N-charr, which consume gammaruses using their teetharmed bones as pincers .
These head morphology transformations allow N2-and N3-morphs to grab gammaruses more effectively. Being a chasing predator L-morph, is characterized by a cone-shaped head armed with very long jaws necessary to catch kokanee in the water column. The transformations in the caudal peduncle length and width also strongly correlate to the feeding strategy reflecting the various swimming abilities of different morphs (Robinson & Parsons, 2002).
The definitive fish length and the length at maturation also correspond to the lifestyle. The smallest size parameters were observed for deepwater morphs while the biggest ones for L-and W-morphs were both associated with predatory feeding. At the second step of diversification the deepwater, lacustrine spawning protomorph would have gone through a diversification in the pelago-benthic resource axes; as a result, two distinct groups, namely B-and S-morphs, may have appeared. Our data demonstrate that they both inhabit the deepwater zone and the isolation is provided by the difference in the place and the time of spawning. At the same step of diversification, the riverine spawning charrs might have split into three groups: L-, W-, and N-morphs. The primary driving factors that lead to the diversification in river-lake charrs seem to be most logically associated with spawning in different sections of the lake tributaries. The isolation during spawning was associated with different riverbed gradients and determined by the critical current velocity hypothetically passable for each morph.
The third step of the evolutionary model we put forward is currently going on and is manifested in diversification of shallow-water Ncharrs into three more groups. The isolation during the reproduction period is ensured using the spawning grounds in different remote tributaries of the lake. Although these groups are very similar in feeding biology, they exhibit significant differences in a head shape.
Our morphological and ecological findings support that the polymorphism of Dolly Varden arising in the Lake Kronotskoe can be considered as a highly probable outcome of the multidimensional disruptive natural selection proceeding within a single ecosystem. Notably, the sympatric diversity found in Lake Kronotskoe is exceptional and exceeds any other known cases for genus Salvelinus in general. It is well acknowledged that any kind of evolutionary processes in sympatry is based on the balance between the effective feeding and successful reproduction strategies (Schluter, 2001). Sympatric speciation is also impossible without reproductive isolation, which can be realized via spatial, temporal, or behavioral segregation (Bolnick & Fitzpatrick, 2007;Seehausen & Wagner, 2014). The morphs in Lake Kronotskoe exhibit the profound level of trophic-based specialization in head and body shape which corresponds to their foraging adaptations and distribution within the lake during growth and maturation. Both the spatial and the temporal isolation are comparatively clear during the reproduction period thus implying that the isolation exists among the morphs.
Food chains in Lake Kronotskoe consist of quite typical units that are common for many dimictic lakes in Holarctic (Jónasson, 1992;Whittaker & Fairbanks, 1958). Thus, the diversity of food objects accounting for the variety of food niches cannot be the only factor that promotes the charr diversification along the different pathways in Lake Kronotskoe. At the same time, the riverine spawning and the preadaptation to the riverine conditions is one of the most important features of Dolly Varden biology. The riverine spawning could have shaped the diversification processes and made the morphs more elaborated and diverse.
The charr community in the deepwater zone in its basic features (feeding and spawning) is similar to the numerous cases of Arctic charr lacustrine populations (Adams, Wilson, & Ferguson, 2008;Alekseyev et al., 2002;Jonsson & Jonsson, 2001;Klemetsen et al., 1997). The only important distinction is that in opposition to numerous pelagic morphs observed in sympatric complexes of Arctic charr the S-morph from Lake Kronotskoe is omnivorous, but not planktivorous which is due to the low abundance of zooplankton in the profundal zone.
Moreover, Dolly Varden in general is hardly able to feed on zooplankton due to low-number and short-length gill rakers in comparison with Arctic charr.
We examined all the abundant morphs of the Lake Kronotskoe charrs inhabiting all ecological niches available in the ecosystem with the exception of planktivorous specialization. All morphs are distant from each other in morphology, distribution across the lake, feeding and spawning strategies. The evidence on the Lake Kronotskoe charrs shows that the morphs are stable from generation to generation in the F I G U R E 5 Three steps of Dolly Varden evolution in Lake Kronotskoe given environment. The possible available ecological niches were comprehensively studied allowing us to suggest that the initial description of charr diversity in the basin is complete. Interestingly enough, we have not found "Riverine Dolly Varden" and "Dwarf charr", which were previously described for this ecosystem (Ostberg et al., 2009;Pavlov, Pivovarov, & Ostberg, 2012;Pavlov et al., 2013). There were no criteria suggested to separate W-morph and Riverine Dolly Varden in those issues except for the habitat in the river mouths. Consequently, it can be identified as W-morph that feeds in river mouths or migrates upstream following kokanee for fattening during the spawning period. The second morph was described from the head of the Kronotskaya River, but the description was performed on four individuals only. Thereby, to associate Dwarf-morph with any of the seven morphs seems to be rather questionable. No more fish with the corresponding characteristics of the Dwarf charr were caught in that area, spawning in the river source was assumed for the Dwarf morph .
The Dwarf charr has recently been found to resemble S-morph by the body shape, but is distinct from the latter by the habitat it dwells in (Markevich, Esin, Saltykova, et al., 2017).
The analysis of mitochondrial DNA polymorphism previously showed a low level of differences between W-, L-, and N-morphs (Ostberg et al., 2009;Radchenko, Salmenkova, & Omel'chenko, 2006;Salmenkova et al., 2005;Senchukova et al., 2012). In this case, the confirmation of reproductive isolation could be analyzed thought the polymorphism of the microsatellite loci and the disruptive selection efficiency through the histo-compatibility complex. However, it is beyond the scope of the present study and will be discussed in the succeeding papers. Furthermore, it should be noted that the recently discovered morphs (B, S, N2 and N3) have never been analyzed by the molecular methods.

ACKNOWLEDGMENTS
The authors would like to express their gratitude to Rune Knudsen 18-14-00319. All parts of this work were performed by all three coauthors in close collaboration. All three authors meet the criteria recommended by ICMJE. We are all participated in the field work, field material processing as well as the text preparation. Both co-authors have been working in the field as well and were processing parts of materials (statistical analysis and spawning ground description). All the authors of this study were actively participating in writing text and preparing the manuscript for publication as well as took part in revising it. There are no any disagreements in results interpretation between authors; we all are accountable for all aspects of this work.
The authors declare no any conflicts of interests.