Intersectional polyphyly of the puzzling Scapania obcordata (Marchantiidae) suggests convergent evolution: resurrection and European occurrence of S. jensenii

Abstract We present an example of morphological convergence among two species of Scapania that were considered to be conspecific. We resurrect S. jensenii from synonymy with the morphologically and molecularly variable S. obcordata because our molecular data show that they are distantly related. Chloroplast trnL-trnF and nuclear ITS sequences resolved S. obcordata in sect. Curtae and S. jensenii in sect. Apiculatae sister to S. obscura and we detected morphological differences in secondary pigmentation and leaf anatomy. The similarity comprises the absence of a pronounced keel in vegetative leaves and a weakly defined stem cortex. These features are virtually unknown in other species of Scapania and probably result from convergence in adaptation to the same environmental conditions, as both taxa inhabit alluvial plains characterised by regular disturbance by flooding and covering with sand and silt. While S. obcordata is widespread in the northern Holarctic and known from the Antarctic, S. jensenii is known from a few localities in Greenland, mainland Norway, the Chukchi Peninsula and the Swiss Alps. Its rarity, the absence of female plants, and the lack of genetic variability suggest that S. jensenii underwent a bottleneck event.


Introduction
The high phenotypic plasticity and a limited set of morphological characters have challenged taxonomists to construct the systematics of bryophytes in the pre-molecular aera.The phylogenetic relevance of different character states has been intensively debated and different classifications have been proposed across essentially all taxonomic ranks (e.g. for liverworts, cf.Gottsche et al. 1844;Stephani 1898Stephani -1924;;Müller 1951Müller -1958;;Grolle 1983;Söderström et al. 2016).Rather unsurprisingly, the application of molecular methods then led to a considerable reshuffling of taxonomic units including the placement of the three major clades, hornworts, liverworts and mosses (Leebens-Mack et al. 2019).Furthermore, the backbone phylogeny of divisions, classes, and orders has been largely clarified in the last couple of years (e.g.Villarreal et al. 2016;Flores et al. 2018;Liu et al. 2019;Yu et al. 2020;Dong et al. 2022).However, many uncertainties remain at lower taxonomic ranks, especially at the level of genera and species and their treatment is constrained by the availability of genetic resources.For many (currently accepted) taxa no single genetic barcode is available and for taxa with available sequences, the sampling of specimens is most often not representative of the geographic distribution and morphological variability within the taxa.Here we uncover a polyphyly in the liverwort genus Scapania (Dumort.)Dumort., nom.cons.that relies on the assumption of the synapomorphic origin of rare morphological characteristics.
In Scapania, the most prominent trait that characterises the genus is the conduplication of the bilobed leaves, i.e. the bend between the leaf lobes usually forms a distinct keel.A few taxa with rather canaliculate (gutter-like) leaves which additionally share a weakly developed stem cortex are currently referred to as a single polymorphic species, S. obcordata (Berggr.)S.W.Arnell (Söderström et al. 2016;Hodgetts et al. 2020).Earlier, Schuster (1974) and Schuster and Damsholt (1974), who extensively studied Scapania plants that share canaliculate leaves and a weakly defined stem cortex suggested that the high morphological variability observed among such morphs has a genetic basis.They accounted for this by describing and recognising different taxa and merged them in the presumably early branching subgenus Jensenia S.W.Arnell (Schuster 1974;Schuster and Damsholt 1974).However, at the same time, they stressed the difficulty to formulate a sound infrasubgeneric taxonomical concept due to high phenotypic plasticity and overlapping character states of the candidate taxa.Jensenia was first established by Arnell (1956) to accommodate S. lapponica (Arnell & C.E.O.Jensen) Steph., which, like S. obcordata, lacks a distinct keel in vegetative leaves.Schuster (1974) and Schuster and Damsholt (1974) also included S. obcordata, S. paradoxa R.M.Schust.and S. perssonii R.M.Schust.and treated Lophozia violascens Bryhn & Kaal., S. lapponica and S. jensenii as synonyms of S. obcordata.Scapania jensenii was interpreted as a weak or juvenile form, although the types are fertile and lack acute leaf lobes characteristic of juvenile plants in the genus.
Subsequently, based on a morphological re-examination, Potemkin (1999) followed by Damsholt (2009Damsholt ( , 2013) ) regarded S. paradoxa synonymous with S. obcordata and S. perssonii synonymous with S. curta var.grandiretis R.M.Schust.and included the subgenus Jensenia in sect.Curtae (Müll.Frib.)H.Buch (Potemkin 2002).Potemkin's view was later supported by the first comprehensive molecular phylogeny of the genus (Heinrichs et al. 2012), in which two accessions of S. obcordata from Svalbard were resolved in sect.Curtae.Consequently, of the taxa assigned to Jensenia by Schuster and Damsholt (1974) only S. obcordata was accepted in the worldwide checklist of liverworts (Söderström et al. 2016).Scapania obcordata was described from Svalbard (Berggren 1875) and reported from arctic and boreal regions in the Holarctic and from the Antarctic (Schuster 1974;Bednarek-Ochyra et al. 2000;Damsholt 2002).The unkeeled leaves and often more or less equal sized leaf lobes make plants of S. obcordata often look more like a Lophozia (Dumort.)Dumort.("lophozioid" plants; Schuster and Damsholt 1974), although well developed and especially fertile female plants often form an indistinct keel and thus show a habit similar to other Scapania species ("scapanioid" plants).
Recent work on Scapania in Switzerland revealed such lophozioid plants in herbarium specimens collected in the beginning of the twentieth century.However, these samples differed in some characteristics such as secondary pigmentation from specimens of S. obcordata from Svalbard.Furthermore, the plants from the Alps would represent geographically disjunct occurrences, with regard to the nearest records of S. obcordata in Scandinavia and this prevented us from pragmatically assigning them to S. obcordata.The lophozioid Scapania plants were represented by three specimens in Z + ZT labelled S. obscura (Arnell & C.E.O.Jensen) Schiffn.They were collected by P. Culmann who recognised their morphological peculiarity, because he initially named them S. lapponica.The specimens were collected at two localities now being submerged in artificial lakes in Bernese Oberland (Grimsel and Räterichsboden reservoirs).Hence, these populations are extinct but studying old maps, we could infer their habitats as siliceous alluvial plains.Based on this we visited similar habitats in the region and encountered extant populations of the taxon at different localities in the cantons of Bern and Uri.Sampling this populations allowed us to conduct preliminary molecular analyses which indicated that sequences of the plants from the Alps were essentially unrelated to sequences of S. obcordata provided by Heinrichs et al. (2012).Consequently, to clarify the identity of the plants from the Alps, we undertook a re-examination of the taxa corresponding to the morphological concept of Jensenia and of specimens assigned to S. obcordata in previous studies.

Taxon sampling
The sampling for molecular analyses included four accessions of the plants from the Alps with Jensenia morphology from Switzerland, two accessions of S. obscura (Norway, Switzerland) and six accessions assigned to S. obcordata, of which fife from Svalbard and one from Siberia (Yamalo-Nenets Autonomous Okrug; Table 1).Three of the accessions of S. obcordata were from the two specimens that were analysed in earlier studies (Konstantinova 123-1-04, KPABG; Vilnet et al. 2010;Hentschel Bryo 0389, GOET;Heinrichs et al. 2012).We re-examined these two specimens to ascertain that our molecular and morphological diagnoses refer to the same material.The specimen collected by Hentschel was split and the duplicate used by Heinrichs et al. (2012) and us for DNA extraction is located in JE.We found that this specimen is heterogenous and includes at least to morphological entities that we considered as elements 1 and 2. Both were morphologically assignable to S. obcordata.Element 1 was characterised by brown-purple colouration, more intensively coloured marginal leaf cells, that were mostly >20 µm wide on the ventral lobe, median leaf cells with moderately thickened walls, concave corner thickening and subangular lumina.Element 2 was characterised by grey-purple colouration, the marginal leaf cell similarly coloured as other leaf cells, ±20 µm wide on the ventral lobe, median leaf cells with irregularly and rather strongly thickened walls, ±rounded lumina and p.p. convex corner thickenings.
The morphological examination comprised a larger set of specimens.Of the plants with Jensenia morphology in the Alps we additionally examined the three specimens collected by P. Culmann at Unteraar (today Grimsel reservoir) and Räterichsboden (today Räterichsboden reservoir), and 26 recent collections from a total of 5 localities in the Central Alps of Switzerland (33 specimens in total; Table 1).Furthermore, we studied c 200 specimens assigned to S. obcordata from Alaska, Antarctic, Canada, Greenland, Russia, mainland Norway and Svalbard from the collections of C, E, LU, TRH, UPC and UPS, including syntypes of S. obcordata (Sarcoscyphus obcordatus Berggr., Insulae Spetsbergenses: Prins Charles Forelands Sund; Adventbay; Brandewijnebay; Mount Misery, S. Berggren, 1868; UPS-B724100-3), the type of S. lapponica (Martinellius lapponicus Arnell & C.E.O.Jensen; Lapponia lulensis, regio Sarjekensis, Kåtokjokk, C. Jensen et H. Wilh. Arnell., 22.07.1902;LD-1431373), the types of S. jensenii (Jungermannia globulifera C.E.O.Jensen; C-M36830-2) and the specimens collected by K. Damsholt near the type localities of S. jensenii (C: KD 69-140, 69-141;Schuster 1974;Schuster and Damsholt 1974).We could not trace the lectotype of S. obcordata mentioned by Schuster (1974) without indicating the herbarium and an isolectotype in S is currently unavailable due to construction work.To locate specimens and records of S. jensenii we furthermore undertook an in-depth literature search.

DNA sampling and laboratory protocols
Following earlier work on the genus (Vilnet et al. 2010;Heinrichs et al. 2012;Potemkin et al. 2021) we choose to sequence the nuclear ITS region (18 S rRNA gene, partial sequence; internal transcribed spacer 1; 5.8 S rRNA gene; internal transcribed spacer 2; 25 S rRNA gene, partial sequence) and the chloroplast marker trnL-trnF including partial sequences of the trnL-and trnF-genes and the spacer in-between.Total genomic DNA was extracted using the NaOH Method (Werner et al. 2002).Amplification of the target regions proved to be challenging and we finally succeeded to generate amplicons by adding bovine serum albumin (BSA) to the PCR solution.The solution comprised 0.6 µL DNA solution, 5 µL DreamTaq PCR Master Mix (Thermo Fisher, USA), 2 µL of 4 µg/mL BSA solution and 1.2 µL of each primer (2.5 pmol/µL).For the amplification of the ITS region we first tried to amplify the whole segment and in cases this was not successful we targeted the ITS1 and ITS2 regions separately by additionally using internal primers.We provide the primers and temperature cycling used for amplification in Table 2.For one specimens of S. obcordata (TRH-B8772, Table 1) amplification of both loci failed.Amplification products were cleaned with a mixture of one unit exonuclease I (20 U/µL; Thermo Fisher, USA) and two units alkaline phosphatase (1 U/µL; Thermo Fisher, USA) and purified fragments were commercially sequenced using the amplification primers by Eurofins Scientific SE, Luxembourg.We then generated the alignment using the online interface of MAFFT v7 (Katoh and Standley 2013) with the E-INS-i aligning strategy and otherwise default settings followed by manual corrections and we scored indels using simple coding (Simmons and Ochoterena 2000).To get an overview of the phylogenetic placement of our accessions, in an initial analysis we performed Bayesian inference (BI) using MrBayes v.3.2.7a (Ronquist et al. 2012) on this alignment.We specified a GTR + G + I model, a sample frequency of 100, stop rule set to yes with critical value for the topological convergence diagnostic (stopval) set to 0.01 and default settings for all other parameters and checked convergence of the runs using Tracer v1.7.2 (Rambaut et al. 2018).For the definitive analyses, we then focused on the sections and subspecies in which our accessions were resolved and reduced the complexity of the data set by limiting the number of Scapania accessions per species, section and subspecies sensu Váňa et al. (2012)  closely related to either accessions of plants with Jensenia morphology from the Alps or of S. obcordata according to the initial analysis.Furthermore, we retained up to two representatives of other species of sect.Curtae and one representative of each other section and subsection.For this reduced dataset we then scored indels and performed BI with the same settings as in the initial analysis and additionally run a maximum likelihood (ML) analysis using RAxML v. 8.2.12 (Stamatakis 2014).For this analysis we also specified the GTR + G + I model and stopped bootstrap analysis automatically using the autoMRE command.Support for the nodes of the best scoring tree out of 50 independent ML runs was assessed using the thorough bootstrapping algorithm with the extended majority rule bootstopping criterion.We summarised the support of nodes from BI and ML analyses using TreeGraph 2 (Stöver and Müller 2010) displaying the BI topology.

Morphology
We found that the plants with Jensenia morphology from the Alps differ in several characters from type specimens of S. obcordata and S. lapponica and other specimens assigned to S. obcordata from the Arctic (Table 3).Most strikingly, purplish secondary pigmentations are absent in all Alpine plants and these plants never develop scapanioid morphs.In contrast, the plants from the Alps were inseparable from type material of Scapania jensenii and are in the following assigned to this taxon (Table 1)  specimens of S. jensenii bear antheridia and female plants where neither detected by Jensen (1906) nor by us in any of the known collections of S. jensenii (Table 1).
The revision of herbarium specimens assigned to S. obcordata revealed four more specimens of S. jensenii from Greenland (additionally to the types, Table 1.) and one from mainland Norway (Nordland, Mo i Rana).The latter was collected by , Table 1) and annotated as "Material identical with that of Lophozia jensenii K.M. from E-Greenland".Damsholt (2002Damsholt ( , 2009) ) published a drawing of a shoot from this specimen in Figure 3 of plate 114 and interpret it as "Juvenile gemmiparous Lophozia-like shoot" of S. obcordata.All other specimens represented either S. obcordata or other species of the genus.
We furthermore detected two records in the literature that we assign to S. jensenii.Schljakov (1973Schljakov ( , 1975) ) reports the species from the Chukchi Peninsula and a further locality in Greenland (Cape Brewster).The drawing provided in Schljakov (1973) sub S. globulifera (С.Jens.)Schljak.coincides with the morphological concept of the species presented here.

Phylogenetic analysis
ITS and trnL-trnF sequences of all four accessions of S. jensenii from the Alps were identical.For both markers, BI and ML analyses resolved these accessions in a supported clade (ITS: PP 1/BS 100, trnL-trnF: 1/86) in sect.Apiculatae H.Buch of subsp.Scapania (sensu Váňa et al. 2012) and sister to accessions of S. obscura (Figures 1 and 2).The accessions of S. obcordata were resolved in a supported clade (0.94/-, 0.90/79) in sect.Curtae of subsp.Scapania and sister to the accessions of S. obscura, S. irrigua, S. helvetica Gottsche and S. curta retrieved from Genbank.The ITS sequences of S. obcordata were variable (Figure 1) and differed from each other in up to eight unique point mutations.Only two accessions (NOS 1 and 4) which were collected at two localities near Adventalen, distant c 30 km from each other had identical ITS sequences and shared one unique point mutation and one indel with respect to the other accessions of S. obcordata.The two elements in the specimen collected by Hentschel differed from each other in one point mutation and from the earlier generated sequence of this specimen in eight (element 1) and seven (element 2) point mutations respectively.trnL-trnF sequences of S. obcordata were identical, except for one of the earlier generated sequences (NO 2, Genbank accession no.EU791626; Figure 2).(2021).Besides confirming the phylogenetic position of S. obcordata in sect.Curtae our data indicate that S. jensenii, which was considered conspecific with S. obcordata, is a distinct species and belongs to sect.Apiculatae.The morphological distinctness of Scapania jensenii has already been recognised by Schljakov (1975), who suggested that this taxon is not closely related to S. obcordata but his work has not been considered in subsequent taxonomical treatments (e.g.Damsholt 2002Damsholt , 2009Damsholt , 2013;;Söderström et al. 2016).Our data furthermore suggest that S. obscura also belongs to sect.Apiculatae.This species was lastly assigned to sect.Scapania but attempts to generate DNA sequences and to prove this molecularly were hitherto largely unsuccessful (Potemkin et al. 2021).In our analyses of ITS and trnL-trnF sequences of specimens from Norway and Switzerland, S. obscura was resolved as sister to S. jensenii.This sister position is paralleled by shared morphological (see 'Differentiation' in the taxonomy section) and ecological characteristics because both species occur in similar habitats at streams.At one of the localities, (Sidelenbach, Table 1) they colonised the same floodplain.

Additions to the phylogeny of Scapania
Regarding the phylogenetic position of S. obscura our results are in conflict with one accession from the Magadan Prov. in Russia available on Genbank (trnL-trnF: JX630058, ITS: JX629927) which was resolved in sect.Curtae in the analysis of ITS sequence data and in sect.Scapania in trnL-trnF (Figures 1 and 2).We aimed at studying the specimen morphologically, but we were unsuccessful in obtaining it.

Polyphyly of S. obcordata and morphological convergence
We show that S. obcordata as previously defined morphologically (Schuster 1974;Schumacker and Váňa 2005) is polyphyletic.Hence, the characteristics formerly used to define the species and the subgenus Jensenia, i.e. not or weakly carinate leaves and a weakly defined stem cortex, either developed independently in sect.Curtae and Apiculatae or represent a preserved ancestral state.The latter scenario is unlikely, because molecular evidence indicates, that the subgenus Scapania (which includes the sect.Curtae and Apiculatae) is a derived lineage (Heinrichs et al. 2012) within the genus and these characteristics are neither shared by other members of the subgenus nor by earlier diverging taxa of the genus.However, it is agreed, that the genus itself most likely originated from a lophozioid ancestor (Potemkin 2002;Vilnet et al. 2010;Heinrichs et al. 2012).Hence, the morphological peculiarities of S. obcordata and S. jensenii seem to represent convergent reversals to the ancestral state, possibly in adaptation to the habitat.Both taxa occur in harsh environments in the arctic or the alpine or subalpine zones, often in the vicinity to glaciers.The sites are periodically disturbed by flooding and overwhelmed with sand or silt.Under these conditions, vertical growth, which is the most common expression of S. jensenii, especially at sites where it is abundant, seems advantageous to emerge from the sand as quickly as possible and the reduction of keels and formation of julaceous shoots could prevent mechanical damage through flooding with the suspension of water and glacial abrasive material.Accordingly, these habitats are commonly colonised by acrocarpous mosses characterised by vertical growth such as Pohlia filum (Schimp.)Mårtensson, Bryum sp.pl., Aongstroemia longipes (Sommerf.)Bruch & Schimp., Philonotis seriata Mitt.or Ditrichum pusillum (Hedw.)Hampe.Among these, P. filum and A. longipes, which are often abundant in these habitats, similarly to S. jensenii and S. obcordata form julaceous upright shoots.The two latter species also share a weakly defined stem cortex.This could be an adaptation that allows accelerated apical growth due to lower investment in thickened cell walls in order to emerge from deposited sand and silt.Schuster and Damsholt (1974) extensively studied the complex of subgenus Jensenia and concluded that it is probably diverse but taxonomically challenging due to an extremely high plasticity and overlapping characteristics.Although we could sort out one of the entities, taxonomic uncertainty persists for the morphotypes that remain under the name of S. obcordata.We observed a high morphological diversity among and within specimens and our molecular data indicate that in fact it coincides with genetic differentiation.For instance, we found that two morphologically distinguishable entities that we encountered intermixed in the specimen collected by Hentschel (Table 1) also differ molecularly.However, an extended sampling is needed to clarify whether the morphological and molecular diversity in S. obcordata relies on reproductively isolated entities and if further taxa need to be recognised.Furthermore, recent collections of S. jensenii from Greenland are needed to address the molecular affinities to the population in the Alps.We could not find any collections of S. jensenii in official herbaria that could be sequenced.

Morphological and genetic variability, sex ratio
The earlier interpretation of S. jensenii being conspecific with S. obcordata (Schuster 1974;Schuster and Damsholt 1974) is probably related to the poor type material of S. jensenii, which can easily be interpreted as a depauperate morph of other species.In fact, the innovation shoots of many Scapania species have a similar morphology as S. jensenii, especially unkeeled leaves (see taxonomy section).Furthermore, the finding of S. obcordata (Damsholt 69-141 and 69-142;Schuster 1974) near the type locality of S. jensenii probably has contributed to the species being taken as a synonym of the former.However, the site where Damsholt collected S. obcordata is about 20 km away from the two sites where Krŭŭse collected the type material of S. jensenii.This can be seen from a map, probably drawn by Damsholt, enclosed with the types of S. jensenii in C.
None of the known collections of S. jensenii revealed female sex organs although we carefully examined all available collections from Greenland and sampled several and partly huge populations in the Alps.Likewise, Schljakov (1973Schljakov ( , 1975) ) does not report female plants in the specimens from Cape Brewster (Greenland) and the Chukchi Peninsula.Hence, globally only male plants are known.This could be explained by an extreme case of sex bias or low expression of female sex organs.Unequal sex ratios are not uncommon in natural populations of dioicous bryophytes, but they are usually biased towards females (Bisang and Hedenäs 2005;Kiebacher et al. 2019).However, an example of male biased sex ratio has been reported in one species of Scapania, S. undulata (L.) Dumort.(Holá et al. 2014).The alternative explanation is that S. jensenii underwent a serious bottleneck event and that only male plants survived.This has some support from the limited distribution of the species and the molecular data.Although suitable habitats are not uncommon in the Holarctic, S. jensenii is only known from few localities and we did not detect intraspecific genetic variability in S. jensenii compared to intraspecific variability in other species of the genus including S. obcordata (Figure 1; Heinrichs et al. 2012).

Phytogeography and conservation
The occurrence of S. jensenii in Greenland, Norway and the Alps is a further example of the well-known floristic connection between these regions which for some plant species including bryophytes has been explained by similar ecological conditions and recolonisation from shared Pleistocene refugia between the Scandinavian and Central European ice shields or refugia in Southern Europe (Dierssen 2001;Ehrich et al. 2007;Hedenäs 2010;Eidesen et al. 2013).However, for another group of species in Greenland genetic data indicate a closer relation to North America than to Europe (Skrede et al. 2006;Eidesen et al. 2013).As we were able to obtain and study only few specimens of S. obcordata from North America, it seems worthwhile to critically revise further specimens from this region which may at least partly represent S. jensenii.Occurrences of S. jensenii in North America, could also explain the hitherto highly disjunct record from the Chukchi Peninsula through connections via the Beringia bridge (Eidesen et al. 2013;Wen et al. 2016).
The fact that from the c 200 specimens of S. obcordata revised only a few represented S. jensenii and only one was not from Greenland, suggests that S. jensenii is extremely rare and that the population in the Alps is important for the conservation of the species at the global scale.Its conservation should therefore be prioritised by the local authorities.The habitats of S. jensenii in the Alps, alpine floodplains and glacial streams are inhabited also by other species of conservation concern such as A. longipes or S. obscura (Kiebacher et al. in press).In Switzerland, these habitats are among the most threatened ones (BAFU 2019) and in the European Union they are considered of major conservation concern in the Natura 2000 framework (habitat code 3220; Council of the European Commission 1992).Beside the construction of hydropower plants, climate warming is certainly an immediate threat to the population of S. jensenii in the Alps.We encountered the species in alluvial zones that some decades ago where still glaciated.Natural succession is likely to vanish these occurrences, especially if the periodic disturbance through overwhelming by sand or silt from glacial streams ceases.The usually abundant formation of gemmae seems to enable the speto spread locally also without sexual reproduction, but its absence at presumably suitable sites within a few kilometres of abundant occurrences (e.g. at lake Grimsel and lake Oberaar with respect to the occurrence at Bächlisboden; Table 1) suggest that dispersal is limited.
Finally, our study highlights the importance of natural history collections to document biodiversity loss through time, which is the first step to identify its causes.The two localities where S. jensenii was collected by Culmann (specimens in Z + ZT, Table 1) at the beginning of the twentieth century were destroyed by the construction of hydropower stations.The same applies to another bryophyte species, for instance Bryum cyclophyllum (Schwägr.)Bruch & Schimp.The only record of this species in Switzerland is from a herbarium specimen (LAU) from the nineteenth century and the locality is today submerged in a reservoir.
Note 1: Schuster (1974) indicates Krŭŭse's collection from Kingorsuaq in C as the lectotype of J. globulifera, without explicitly specifying that this represents typification.Since we did not find any earlier typification and since the use of 'designated here' or an equivalent is not required before 2001 (Turland et al. 2018) we interpret this as valid typification.However, in C there are two specimens that correspond to Schuster's typification (on one sheet below the specimen from Sieraq Dal) and we here select the specimen at the bottom of the sheet (C-M36832) as lectotype, because the material is more abundant than the specimen with the barcode C-M36831 in the middle of the sheet.Opposed to this, on the herbarium sheet the specimen from Sieraq Dal (C-M36830) is annotated as "Lectotypŭs".We could not identify an effective publication for this and consequently regard this specimen as syntype.From the three type specimens this latter one comprises the most abundant material.
Note 2: The isolectotype (C-M36831) is annotated as having been collected in 1898.The specimen contains a note that this annotation is erroneous.

Variation
Overall, Scapania jensenii is a rather stenotypic species, but it varies considerably in the size of the shoots which can be very small (down to 0.2 mm wide).This varies between individual patches as well as within patches and also well grown plants are usually composed of shoots of different sizes.In small shoots the leaves are generally more erect (Figure 3).
Secondary brownish colouration is usually present but plants from shaded microsites can be purely green.Furthermore, there is some variation in the shape of the gemmae.In some specimens, including the types, we observed a few distinctly elongate 2-celled gemmae, sometimes slightly constricted (Figure 4).However, the majority of gemmae are almost spherical or at most broadly ellipsoidal.

Differentiation
Essentially all Scapania species when grown under extreme conditions, or their innovation shoots, can look similar to S. jensenii (small shoots with scarcely carinate and subequally lobed leaves).Such morphs can usually be distinguished from S. jensenii by the shape of the lobes and the gemmae.The lobe apices of such plants and shoots are acute in most Scapania species, also in species where mature plants have rounded lobes.In contrast, they are mostly blunt in S. jensenii.The spherical or shortly ellipsoidal gemmae of S. jensenii are only shared by S. obcordata (Figure 5), S. compacta (Roth) Dumort.and S undulata.All other European species have narrowly ellipsoidal, elongate or fusiform gemmae, which are usually much longer than wide.Individual gemmae can, however, also be elongate in S. jensenii, but they are never fusiform (Figure 4, Jensen 1906).The gemmae of S. obcordata are similar in shape to those of S. jensenii but are usually distinctly larger and, in contrast to S. jensenii, may develop reddish or purplish secondary pigments (Figure 5; Table 3).Further differences between these two species are listed in Table 3.
Scapania obscura is closely related to S. jensenii and the two species share the subequal leaf lobes, brownish secondary pigmentation, the lack of reddish pigments and both occur in alpine habitats influenced by water courses over siliceous bedrock.Furthermore, also in S. obscura the keel is scarcely developed but usually present as a rounded angle between the two lobes whereas in S. jensenii it is absent.Furthermore, Scapania obscura is a distinctly larger plant (shoots up to 2 mm wide vs. up to 1.2 mm) and the ventral lobe is long decurrent (vs. at most shortly decurrent).

Ecology
In the Alps, S. jensenii occurs in alluvial zones or on more or less flat banks of glacial streams where it may cover several square metres or grow intermixed between other bryophytes.The sites are periodically flooded with silty water.Vascular plant cover is generally low and often characterised by small-grown alpine willow species, Cyperaceae (Eriophorum scheuchzeri Hoppe, Carex sp.pl.), Epilobium sp.pl. and a more or less random set of forbs and Poaceae that colonise these bare soils.Highly dynamic alluvial areas (disturbed at least once a year) that are often covered exclusively by Pohlia filum as well as rarely disturbed sites with denser vascular plant cover are more rarely colonised.Scapania jensenii seems to depend on constantly moist conditions.It usually grows at most 20 cm above the waterline of adjacent streams or at sites with high groundwater level.The substrate is formed by fine siliceous sand or clay, sometimes covered by a thin humus layer that rarely exceeds 1 cm.

Distribution
Scapania jensenii is to date known from several localities in Greenland and the cantons of Bern and Uri in the Swiss Alps, and from mainland Norway and the west part of the Chukchi Peninsula in easternmost Asia (one record each; Table 1; Figure 6).At two of the localities in Switzerland (Bächlisboden, Steisee/Umpol) the species locally covers several square metres and occurs in an area with an extension of up to 2 km.At the other localities in Switzerland, we observed only a few square centimetres and the population size in Greenland, Norway and Chukchi is unknown.

Figure 1 .
Figure 1.Bayesian inference of nuclear its sequence data.numbers above branches are Bayesian posterior probabilities ≥0.5, numbers below branches are bootstrap support values ≥50 of branches obtained from maximum likelihood analysis of the same dataset.accessions of Scapania jensenii and S. obcordata are in bold, accessions retrieved from genBank are prefixed by the accession number.
. The specimens collected by K. Damsholt in the Angmagssalik Distr. of Greenland near the type locality of S. jensenii (C: KD 69-140, KD 69-141;Schuster 1974;Schuster and Damsholt 1974) differ from the types of S. jensenii as well as from the alpine collections in the same characteristics as other specimens of S. obcordata and are assigned to that taxon.In none of the 33 specimens of S. jensenii from the Alps we encountered female plants.Almost all specimens included at least a few antheridia bearing shoots, even when they consisted of poorly developed plants.In well grown plants, shoots with antheridia were usually abundant.Likewise, many shoots of the type

Figure 2 .
Figure 2. Bayesian inference of Plastid trnL-trnF sequence data.numbers above branches are Bayesian posterior probabilities ≥0.5, numbers below branches are bootstrap support values ≥50 of branches obtained from maximum likelihood analysis of the same dataset.accessions of Scapania jensenii and S. obcordata are in bold, accessions retrieved from genBank are prefixed by the accession number.
Our results provide additions to the phylogeny of the genus Scapania presented byHeinrichs et al. (2012) andVáňa et al.

Table 1 .
collection data for all records and specimens of Scapania jensenii and for specimens of S. obcordata and S. obscura examined molecularly: taxon (in brackets taxon according to label); accession code (iso 3166-2 country code and sequential number) used in figures 1 and 2; voucher information including locality, elevation, coordinates, date, collector and collection number (herbarium with barcode if available or reference); genBank accession numbers: nrits, trnL-trnF.

Table 2 .
Primers and thermal cycling protocols used for Pcr to amplify the target loci.