Disentangling Scandinavian species hidden within Meesia uliginosa Hedw. s.l. (Bryophyta, Meesiaceae)

The molecular variation (atpB–rbcL spacer, rpl16 G2 intron, trnGUCC G2 intron) among 50 Scandinavian specimens of Meesia uliginosa s.l. was analysed in the context of three other Meesia species, with Paludella squarrosa as outgroup. The molecular variation within M. uliginosa correlates with morphology, and shows that three species exist: M. uliginosa Hedw., M. minor Brid. and M. minutissima Hedenäs sp. nov. Whereas M. uliginosa s.str. is widespread, in Scandinavia the other two species are restricted to the mountains. The three species differ in quantitative features, such as, plant size, leaf size, costa width, leaf apex shape, seta length and spore capsule size. Because the species overlap in several of these features, a combination of seta length and the ornamentation of the exostome outside are the safest characters for identification. Leaf apex shape, small spore capsules in M. minutissima and large spores in M. minor serve as additional identification help. Meesia minutissima is presently only known from Scandinavia. The other two species are more widespread, although further investigations are required to find out their exact distributions.

Despite more than 200 years of investigations, scientists and gifted amateurs frequently discover bryophyte species new to Europe. Disregarding introductions by man, this includes species found for the first time in Europe and species new to science. Based on recent publications on new species, Mediterranean and oceanic regions (Gallego et al. 2000, Muñoz et al. 2009, Cezón et al. 2010, Bosanquet and Lara 2012, Hedenäs et al. 2014) and the European mountains and far north (Köckinger and Kučera 2007, Draper and Hedenäs 2009, Köckinger et al. 2012, Hedenäs 2018 are likely the areas that harbour most unrecognized species. During many years of fieldwork in northern and mountainous Scandinavia, I have come across several species which variation suggests that they may actually consist of more than one species. In connection with a study of northern and mountain intraspecific genetic variants within widespread species (Hedenäs 2019), it became evident that the genetic variation within Meesia uliginosa Hedw. correlates with morphological differentiation and deserves further study.
Within the Meesiaceae, seven of the thirteen recognized species belong to Meesia, and the remaining ones are distributed among four small genera (Frey and Stech 2009). Meesia occurs predominantly in temperate to Arctic and Antarctic regions (Frey and Stech 2009), and four of the species are circum-Arctic (Frey et al. 2006, Ignatov et al. 2006, Favreau and Brassard 2011: M. hexasticha (Funck) Bruch, M. longiseta Hedw., M. triquetra (L. ex Jolycl.) Ångstr. and M. uliginosa. Some recognized species of Meesia are morphologically variable or intermediate between other species, which has been suggested to be a result of hybridisation, such as M. hexasticha, possibly a hybrid between M. uliginosa and M. triquetra (Nyholm 1998), and M. longiseta × triquetra Lindb. & Arnell (Lindberg and Arnell 1890). Variability of a different nature was found within M. triquetra, where Montagnes and Vitt (1991) showed that there exist latitudinal and ecoclimatic gradients in morphology in North America, and discussed if environmental or genetic variation could explain these differences. For European M. uliginosa, early authors thought that small phenotypes found in mountain areas represent distinct species (Bridel 1803, Bruch 1826, Bridel-Brideri 1827. Also in Scandinavia, such small mountain phenotypes of M. uliginosa were noted early, for example by Hartman (1832) and Ångström (1844), but although such phenotypes are still noted in modern Scandinavian floras, their taxonomic status has not yet been clarified (Nyholm 1958, 1998, Hallingbäck et al. 2008. Meesia uliginosa s.l. occurs almost throughout Scandinavia (Sweden's Virtual Herbarium: <http://herbarium.emg. umu.se/index.html>; Norwegian Biodiversity Information Centre: <www.biodiversity.no/>, accessed 27 June 2019), but is rare or has vanished from many southern areas. The species, in its wide sense, is extremely variable in size as well as in some other characters. In the lowlands it typically grows in rich fens, the plants are large, up to ca 30 mm tall, the leaves have rounded apices, and the seta may be up to 70 mm tall. In exposed habitats in the mountains, it may in addition grow on base-rich, bare, peaty soil or often in rock crevices, and many plants are considerably smaller. Mountain plants vary in shoot length between 2 and 30 mm, in leaf apex shape from acuminate to rounded, and in seta length between 4 and 70 mm. In a recent study where Swedish M. uliginosa was included, three molecular lineages were revealed (Hedenäs 2019). Here, the wide morphological variation in M. uliginosa is evaluated in detail in the light of these molecular results, to understand whether it is primarily due to phenotypic plasticity or if more than one species could be involved.

Studied material
For the molecular evaluation, 50 Swedish specimens of Meesia uliginosa s.l., were studied. The sampling represents different areas and environments where it occurs in Sweden, and covers its known morphological variation. The sequences for M. uliginosa were available from Hedenäs (2019); new sequences were generated for two specimens of each of the species M. hexasticha, M. longiseta and M. triquetra, to assess the position of M. uliginosa s.l. within the genus, and for Paludella squarrosa (Hedw.) Brid., another member of the Meesiaceae, which was used as outgroup (cf. Goffinet et al. 2001Goffinet et al. , 2004. The molecularly studied specimens are listed in Table 1. The morphological analysis was based on ten specimens representing each of the three molecular entities that were revealed within M. uliginosa s.l. (Table 1).

Molecular methods
The molecular part of this study is based on the plastid atpB-rbcL spacer (atbB-rbcL), rpl16 G2 intron (rpl16) and trnG UCC G2 intron (trnG). The methods used to generate the new sequences included here are described in Hedenäs (2019).

Sequence editing and analysis
Nucleotide sequence fragments were edited and assembled for each DNA region using PhyDE 0.9971 (<www.phyde. de/index.html>, accessed 22 November 2018). The assembled sequences were aligned manually in PhyDE. Regions of partially incomplete data in the beginning and end of the sequences were identified and were excluded from subsequent analyses. Gaps were coded in SeqStat (Müller 2005), using the simple indel coding method of Simmons and Ochoterena (2000), and since they provided additional evidence to distinguish haplotypes they were included in the analyses. The sequence alignments used in the analyses are available on request. GenBank accession numbers are listed in Table 1.
Reticulation was revealed in a preliminary analysis using TCS (Clement et al. 2000; results not shown), and relationships were therefore evaluated in a network context using NeighborNet (NN) split networks produced in SplitsTree 4.12.6 (Huson and Bryant 2006). Potential support for lineages in a tree context was tested by Jacknife analyses (1000 replications) performed with the program TNT (Goloboff et al. 2003).

Morphological study and analysis of measurements
After the molecular relationships among the studied M. uliginosa s.l. specimens had been clarified, the morphology of ten selected specimens from each of the three distinguished entities was studied. For one of the entities, four specimens were included in the molecular sample, but since seta length in combination with exostome ornamentation unambiguously distinguish this entity, another six specimens could be added for the detailed morphological comparisons. Recent treatments based only on morphology (Nyholm 1958, 1998, Hallingbäck et al. 2008) had failed to distinguish well-circumscribed entities within M. uliginosa s.l. Therefore, both standard comparisons of qualitative and quantitative characters and detailed measurements of selected gametophyte and sporophyte features were performed, employing dissecting and compound microscopes.
Specimens for which selected gametophyte and sporophyte features were measured in detail are indicated with an asterisk (*) in Table 1. For each of these specimens, three vegetative leaves were sampled from two shoots (two leaves from one stem and one from the other, to avoid sampling all leaves from an untypical shoot for the specimen). In each leaf, the following features were measured or scored. (a) Length and maximal leaf width (mm), (b) costa width near base and 2/3 way up leaf (µm), (c) length (µm), width (µm) and length to width ratio of 20 cells in the middle portion of the upper lamina, and (d) leaf apex shape, scored as acuminate (<45° = 1), acute (45-90° = 2), obtuse (> 90° but still distinctly 'pointed' = 3) or rounded (4). When mature sporophytes were present, (a) length of the capsule neck (mm), (b) length and width of the urn (mm), (c) length (µm), width (µm) and length to width ratio of 20 exothecial cells in the dorsal portion of one arbitrarily selected urn and (d) spore diameters (µm) were measured. Other than for the numbers of exothecial cells in one capsule per specimen with mature sporophytes, the number of sporophyte measurements depended on the number of available sporophytes in good condition. Temporary images of leaves and cells were taken through a microscope using a digital camera and the Olympus cellSens Standard 1.13 software for automatic and continuous image stacking. Measurements were taken from these leaf and cell images, using the Olympus cellSens Standard 1.13 software.
Comparisons of the detailed measurements among the three entities within M. uliginosa s.l. are based on two approaches. First, measurements were compared between the entities. Potential influence of leaf size on lamina cell size was evaluated by adjusting cell sizes to a standard leaf length of 1.5 mm and a width of 0.3 mm, by dividing the actual Table 1. (A) Specimen data and GenBank accession numbers for the sequences; accession numbers of newly generated sequences begin with MN. All specimens are in herbarium S, and except where noted their geographical origin is Sweden. (B) Swedish Meesia minutissima specimens identified by morphology and that were included in the detailed morphological study. * = leaf and selected sporophyte characters were measured in the detailed morphological study; LH = L.Hedenäs; NA = Sequence not available. leaf lengths or widths with these values and multiplying the resulting values with the cell lengths and widths, respectively. Shapiro Wilks W-test (normality) was mostly statistically significant, indicating that the data do not meet the criterion of normality. Thus, the nonparametric Kruskal-Wallis test for multiple comparisons was used to compare the measurements among or between the entities, respectively. Second, the gametophyte and sporophyte measurements were subjected to separate principal component analyses (PCA) to see whether the combined information within each data set corresponds with the molecularly identified entities. For the PCA the mean values for lamina and exothecial cell sizes were used to represent each leaf or capsule urn, and mean spore diameter to represent spore size in a specimen. For the gametophyte, leaf length, width, apex shape, costa width at the two positions, and the mean lamina cell length, width and cell length to width ratio, in total eight parameters, were included. For the sporophyte, mean capsule neck length, urn length and width, and the mean exothecial cell length, width and cell length to width ratio, and mean spore size, in total seven parameters, were included. All statistical calculations were made in STATISTICA 12 (StatSoft 2013).

Geographical distributions
The geographical distributions of the two species that are here segregated from M. uliginosa were evaluated based on a selection of specimens present in the Swedish Museum of Natural History (S) and the Museum of Evolution in Uppsala (UPS). Detailed information on these specimens is available at the Swedish Virtual Herbarium (<http://herbarium.emg.umu. se/index.html>, accessed 19 August 2019).

Morphological evaluation
When sporophytes are present, the three species within Meesia uliginosa s.l. can be distinguished from each other by a combination of seta length and the ornamentation of the exostome outside (Fig. 2). The seta in M. uliginosa s.str. is 17-70 mm and the exostome outside is smooth, or faintly reticulate to striolate. In M. minor the seta is 8-37(46) mm tall and the exostome outside has a well-developed reticulate or partly irregularly cross-striolate ornamentation, and in M. minutissima the seta is 4-17 mm tall and the exostome outside is almost smooth, or faintly cross-striolate to obliquely striolate. In addition, the apex of well-developed vegetative leaves is mostly rounded or obtuse in M. uliginosa  (Fig. 3). Meesia minor overlaps to some degree with both the other species in the gametophyte features but hardly in the sporophyte features, whereas the two other species are distinct from each other in the PCAs. For the gametophyte, leaf cell length to width ratio and, to some degree, length contribute to the distribution along the second axis, whereas the other features correlate with the first axis (Fig. 3A, left). For the sporophyte, exothecial cell length to width ratio and, to some degree, cell length and spore size contribute to the distribution along the second axis, whereas all other measured features correlate with the first axis (Fig. 3B, left).
All three species within M. uliginosa s.l. differ from each other in seven of the 18 individual quantitative characters measured in detail. Meesia uliginosa differs from the other two species in four additional characters, M minor from the two other species in two additional ones, and M. minutissima from the other two in four (Table 2). When cell sizes were adjusted to a leaf length of 1.5 mm and a leaf width of 0.3 mm, the cells of M. uliginosa appear smaller than those of the other species, even if they are longer than those of M. minutissima and wider than those of both the other species in the actual measurement. For several characters, there is a wide overlap between the species despite statistically significant differences (Fig. 4). Additional observed differences are mentioned in the key and species descriptions.

Geographical distribution and habitat
The known Scandinavian distributions of Meesia minor and M. minutissima are mapped in Fig. 5, based on examined specimens in S and UPS. Whereas M. uliginosa s.str. occurs both in lowland and mountain environments, both the other species are typical mountain ones, with the somewhat more common M. minor collected between 460 and 1350 m a.s.l. and M. minutissima between 550 and 1400 m a.s.l. These two species can grow on exposed soil as well as in rock crevices and occur mainly in areas rich in calcareous or at least somewhat base-rich bedrock. Meesia minutissima appears to grow at on the average more exposed sites than M. minor, but otherwise there does not seem to be any clear distinction between the habitats of the two species. On two occasions, the two were even collected within the same square decimetre.

Discussion
The three entities of Meesia uliginosa s.l. belong to three separate and well-supported molecular lineages, with partly different affinities to other Meesia species. They differ in several morphological features, and therefore there can be no doubt that they represent three species despite their somewhat similar appearances. Any other treatment would be inconsistent as long as we recognize M. triquetra and M. longiseta as distinct species. Among the three entities segregated from M. uliginosa s.l. (Fig. 1), M. uliginosa s.str. clusters with M. longiseta and one of the M. hexasticha accessions, whereas the two well-supported lineages corresponding with M. minor and M. minutissima, respectively, are more closely related to M. triquetra and the second accession of M. hexasticha. Meesia minor, M. minutissima and M. uliginosa display statistically significant differences in several quantitative morphological characters, and can be efficiently distinguished by a combination of seta length and the ornamentation of the exostome outside. Additional important distinguishing characters are the shape of the leaf apex, the capsule size and spore size (Table 2, Fig. 4).
The situation for M. hexasticha is different. One of the markers for its accession D1500 could not be generated, but high support for the positions of this accession and D1499 in different portions of the NN split network suggest that this species is not homogeneous. This would agree with a hybrid origin (Nyholm 1998) and, if this hypothesis is correct, suggests that hybridisation has occurred at least twice. If, as suggested by Nyholm (1998), its parental species are M. uliginosa and M. triquetra, then M. uliginosa is the maternal parent for D1499 whereas M. triquetra is the maternal parent for D1500, since chloroplasts are maternally inherited in mosses (Duckett et al. 1983, McDaniel et al. 2007, Natcheva and Cronberg 2007. To clarify the origin of different M. hexasticha populations, additional material of M. hexasticha must be studied and a nuclear marker should be included in the evaluation. Meesia hexasticha is red-listed in several European countries (Ştefănuţ and Goia 2012, Henriksen and Hilmo 2015, Westling 2015, Hyvärinen et al. 2019. However, if the name M. hexasticha represents a plant phenotype that originated repeatedly due to hybridisation between other Meesia species, its inclusion in red-lists as an independently evolving lineage at the species level is questionable. Investigations during the last couple of decades have added many moss species to the flora of the Scandinavian mountain range and far north. Besides Meesia, examples can be found in several genera, including Oncophorus (Hedenäs 2017), Sanionia (Hedenäs 1989) and Schistidium (Blom 1996) species new to science and species raised from lower taxonomic ranks in Drepanocladus (Hedenäs 1992) and Schistidium (Blom 1996). For other species, their status was earlier unclear, like in Oncophorus (Hedenäs 2017(Hedenäs , 2018, Orthothecium (Hedenäs 1988), Schistidium (Blom 1996) and Sciuro-hypnum Hedenäs 2008, 2009). Finally, in some cases finds were made outside previously known ranges, as in Campylium (Jacobson and Hedenäs 2015), Encalypta (Høitomt et al. 2016), Funaria (Rumsey 1990), Schistidium (Blom 1996), Timmia (Carlsson 2003) and Tortella (Hassel and Høitomt 2013). In addition to this recently discovered species diversity, so far unrecognized species still occur in these areas (Hedenäs unpubl.). In Europe, mountain regions and regions of the far north belong to those that will be most heavily affected by the future temperature increase (Berglöv et al. 2015a, b, Nylén et al. 2015, which underlines how urgent it is to explore northern and mountain diversity. As shown in the modelling study of Moen et al. (2004), most of the Swedish alpine region could potentially disappear until the year 2100. According to the models, such changes will be especially crucial for meadows, heaths and wetlands, where only a few per cent may remain. Kullman (2010) reviewed already visible changes in the Scandinavian mountain vegetation, but doubted that we will see very severe future negative effects of the changing climate. However, even if predictions are always uncertain, the magnitude of the changes suggested by the models of Moen et al. (2004) are too substantial to ignore.

Key to the Scandinavian species of the Meesia uliginosa complex
Note. It is important to study numerous non-perichaetial leaves from the upper portions of shoots to judge the shape of the leaf apex.   Fig. 1), along the first two axes in a PCA. This PCA is based on the length of the capsule neck (CNL), urn length (CAL) and width (CAW), exothecial cell length, width and length/ width ratio (EXL, EXW, EXR), and mean spore size (SPO). Capsule sizes based on 2-5 measured capsules per specimen, exothecial cell sizes and length/width ratios are the mean values of 20 measured cells in an urn, and spore sizes are the mean values of 13-32 measured spores in M. minor, 11-32 in M. minutissima and 10-26 in M. uliginosa s.tr. Factors 1 and 2 explain 57.8% and 23.6% of the variation. hymenostoma Card., sp. Nova. Chili, monte Baguales, T. Halle, 1909' in herb. Roth (lecto-: PC-CARD, n.v. (Matteri and; iso-: H-BROTH, n.v., PC-CARD, n.v., S, B166176!).
This is the largest of the three species within M. uliginosa s.l. The seta length varies between 17 and 70 mm, the exostome outside is smooth or faintly ornamented, and the leaf apices are mostly rounded or obtuse. Specimens with a relatively short seta and obtuse leaf apices could easily be confused with M. minor, but the latter has a lower exostome outside with well-developed ornamentation and on Table 2. Means plus standard errors for character measurements in Meesia uliginosa s.str., M. minor and M. minutissima. When mid-leaf cell length and width were adjusted, these were adjusted to leaves with a length of 1.5 mm and a width of 0.3 mm, respectively. The total number of measurements, n, are indicated for each species (M. uliginosa, M. minor, M. minutissima) under each character. Different letters (a, b, c) appended after the values for a character indicate statistically significant pair-wise differences between species, as revealed by the Kruskal-Wallis test, using the Bonferroni corrected p value corresponding with p < 0.05 (i.e. < 0.00092593).   Fig. 1). For numbers of measurements, see Table 2.

Nomenclatural notes
The lectotype of Meesia uliginosa Hedw. (Ochyra et al. 2008) has a seta that is ca 46 mm long and the outer exostome ornamentation is weak. The vegetative as well as perichaetial leaves have a rounded apex. Thus, the name M. uliginosa Hedw. should be applied to the largest of the three species within M. uliginosa s.l.
The S isotype of Meesia hymenostoma Cardot & Broth. has a faintly ornamented exostome, and clearly belongs to M. uliginosa s.str.

Habitat and known distribution
This species is most frequent in mineral-rich fen or spring habitats, but it also grows on bare, peaty soil, occasionally in escarpments, in base-rich habitats. In Scandinavia, this is the most widespread of the three species within M. uliginosa s.l. (field observations; Hedenäs unpubl.). It occurs both in the mountains and in the lowlands, but is rare or has vanished from large portions of southern Scandinavia. The species is probably widespread in Europe and temperate to Arctic areas of Asia and North America (Nyholm 1998, Xing-jiang and He 2007, Vitt 2014, Ignatov and Ignatova 2018. It occurs in southern South America (type of Meesia hymenostoma Cardot & Broth. and Matteri and Ochyra 1999), whereas reports from the Antarctic area Lewis-Smith 1999, Ochyra et al. 2008) Ochyra and Lewis-Smith (1999)].

Nomenclatural notes
There is only one specimen of Meesia minor Brid. in herbarium Bridel, with label information agreeing with information in the protologue, and this is therefore considered to be the holotype of the name. The stem leaves are gradually acuminate or narrowly obtuse, the seta is 16-18 mm tall, and the exostome outside ornamentation is distinct (Fig. 6A).  , according to information just preceding the paper in the journal, it seems very unlikely that the second G specimen could have been collected and seen by Funck before the publication. The three remaining G and M samples all come from the Swiss Alps, and here the one in G is selected as lectotype since this is the only one that with certainty both agrees with the protologue and was collected before March 1826. Its leaves are acuminate or narrowly obtuse, and the seta is 8-15 mm. Unfortunately, its exostome is in a too poor condition for evaluation. [1]796' (two labels for one specimen), on sheet 2 in folder 621, reg. no. B 31 0621 02], and one in herbarium Hedwig-Schwaegrichen in G ('b)', 'angustifolia, alp Helvet.' s.n., on the sheet with the lectotype of Meesia uliginosa Hedw.). Because all specimens fit the protologue, the well-developed one that is richest in material is selected as lectotype of the name (Fig. 6B).
Ceratodon kinggeorgicus Kanda was synonymised with M. uliginosa by Ochyra and Lewis-Smith (1999). Based on their illustration and since their remark that the Antarctic material is uniform and agrees with the concept of 'var. minor', M. minor is probably the species occurring in the Antarctic.

Habitat and known distribution
This species grows on bare, peaty soil or rocks, or often in rock crevices, in base-rich habitats. It seems to grow in, on the average, somewhat more exposed habitats than M. minor. In Scandinavia, it is widespread in the mountains (Fig. 5B), where field observations (Hedenäs unpubl.) suggest that it is much less common than M. minor. Meesia minutissima is presently only known from Scandinavia, but in view of its widespread habitat it likely occurs also in other mountain regions and in the far north. Because the description of M. uliginosa from Arctic North America by Favreau and Brassard (2011) mentions seta lengths from 6 mm it seems likely that their M. uliginosa includes M. minutissima.