Identification and taxonomy of Betula (Betulaceae) in Great Britain and Ireland

Floras and identification guides typically do not list the full range of identification features that discriminate between Betula pendula Roth and B. pubescens Ehrh., and some accounts contain errors. Discriminatory identification features are described and illustrated. Introgression from the diploid, B. pendula into B. pubescens is widespread. It also occurs from B. nana L. into B. pubescens, but obviously introgressed examples of the latter are probably rare. Betula pubescens presents the greatest unresolved identification difficulties. Trees identified as B. pubescens subsp. celtiberica (Rothm. & Vasc.) Rivas Mart., particularly in Wales, may be part of the variation within British B. pubescens, due to introgression from B. pendula. The most widespread of the small-leaved forms of B. pubescens is correctly named var. fragrans Ashburner & McAll. Several misapplied names have been given to this taxon, most recently subsp. tortuosa (Ledeb.) Nyman, which does not occur in Europe. Not all small-leaved B. pubescens are var. fragrans; some examples show clear signs of introgression from B. nana, and may be referable to the northern var. pumila (L.) Govaerts.


Introduction
The taxonomy and naming of Birch (Betula) has been the source of confusion and dispute ever since Linnaeus published his Species Plantarum in 1753. Linnaeus's Betula alba L. is now accepted as comprising two species, B. pendula Roth (Silver Birch) and B. pubescens Ehrh. (Downy Birch). A third species Betula nana L. (Dwarf Birch) is distinctive, and taxonomically uncontroversial.
In combination, B. pendula and B. pubescens are almost ubiquitous at hectad scale in Great Britain and Ireland. However their frequency relative to each other varies markedly (Fig. 1). B. pubescens is the more frequent species across almost all of Ireland, Scotland (except the east), Wales, north-west and parts of southwest England. B. pendula is the more frequent species in north-east, central and southern England, and in east and north-east Scotland. It is likely that distribution maps for these two species contain some errors (e.g. Worrell & Malcolm, 1998). Problems relating to the identification of British and Irish Betula are due to: 1) Identification features of B. pendula and B. pubescens being poorly and incompletely described in some popular Floras and Field Guides.
2) The occurrence of the F1 hybrid between B. pendula and B. pubescens, Betula x aurata (Borkh.). 3) Introgression from the diploid (2n=28) B. pendula and B. nana, into the tetraploid (2n=56) B. pubescens. 4) The occurrence of distinctive local forms, especially of B. pubescens. 5) Planting of alien (to the local area) genotypes and species, and effects of habitat disturbance.
6) Misapplication of names to taxa within B. pubescens. 7) Misleading interpretations of the relationship between the small-leaved B. pubescens in the Highlands of Scotland and Scandinavian and Icelandic 'mountain birch'. These issues are discussed in turn below.

1) Identification features of B. pendula vs. B. pubescens and of B. nana.
Commonly used Floras and identification guides typically do not list the full range of identification features that discriminate between B. pendula and B. pubescens, and some accounts contain errors. Examining a classic example of the small-leaved B. pubescens subsp. tortuosa (sensu Stace, 2019) in the Highlands of Scotland, a botanist using one of the two popular Field Guides (Rose, 2006;Streeter, 2009) might well be stumped as to what they were looking at. Given that such trees may have whitish bark, pendulous and glabrous fine twigs, that may have frequent glands (or warts), and leaves that may be almost glabrous later in the summer, they may well decide they are looking at a form of B. pendula. The latest edition of the New Flora of the British Isles (Stace, 2019), has some significant errors in the keys, text and illustrations. The key to species in Stace (2019) does satisfactorily distinguish between B. pendula and B. pubescens. However, the drawing of the fruiting bract of B. pendula does not show the diagnostic shape and the drawing of the fruiting bract of B. pubescens is wrong; the fruiting bracts are said to have lateral lobes "directed laterally or slightly backwards" whereas they point forwards (unless introgressed), which is a key difference from B. pendula. The leaves of subsp. tortuosa (sensu Stace, 2019) are not conspicuously hairy as stated in the key to subspecies, and the shoots frequently have many (not few) sessile glands. It is described as usually shrubby, but is almost always a tree. Of easily accessed identification guides, the account in Plant Crib (Rich & Jermy, 1998, and available online) is reliable. For discussion of B. pubescens subsp. celtiberica see Section 4, below.
The Atkinson discriminant function (ADF) (Atkinson & Codling, 1986) is a recommended tool to distinguish between B. pendula and B. pubescens using three leaf characters; details are given in Stace (2019). The function was based on a sample of 104 trees in 14 self-sown populations in England and Scotland. The discriminatory power of the function has recently been tested on a much larger dataset (Wang et al., 2014b). They calculated the ADF for 944 Betula trees (780 B. pubescens and 164 B. pendula) from 105 populations. The species identifications had previously been determined using 12 microsatellite loci (Wang et al, 2014a).
Using an ADF value of zero as the boundary between the two species, as originally proposed, the ADF value correctly identified 96.4% of samples. Using a boundary value of -2 raised the success rate to 97.5%. However, between ADF values of -11 to +3, error rates exceeded 10%. (Fig. 2). The practical usefulness of the ADF seems to be limited. It reliably distinguishes between the two species in cases where leaf shape is fairly obviously one or other species. When it would be most useful is when their leaf shapes are most similar. Here the error rate increases to over 10%. Some of the trees sampled by Wang et al. (2014b) showed evidence of introgression, but they concluded that the morphological continuum between the leaf shapes of the two species "is inherent to the species, and hybridisation is not its major cause". Betula nana (Fig. 3) is a highly distinctive shrub, prostrate or erect (maximum height c.1 metre), with small, shining, dark green, subcoriaceous, rounded leaves with crenate teeth (teeth not acute and triangular as in the other species and hybrids), and densely 'furry' hairy shoots. It has a totally different arrangement of catkins with overwintering male catkins never terminal, but lateral, usually below females, and over winter looking like elongated buds, and erect fruiting catkins. Its hybrid with B. pubescens, B. x intermedia (Fig. 27) is little known to British botanists, having only been recorded by c.12 recorders since 1987. B. nana, recorded from 101 hectads post 1999, is almost restricted to the Scottish Highlands; it is rare in northern England. The hybrid is recorded from 10 hectads post 1999, and was reported from Northumberland in 2020, new to England. (Hectad counts from BSBI Distribution Database, March 2021).  Table 1.
Identifications should always be based on consideration of as many features as possible. I have drawn on a wide range of literature and personal experience in the field (mainly in north Scotland) when compiling this. Note that the name B. pubescens var. fragrans (Ashburner & McAllister) is used in place of B. pubescens subsp. tortuosa (sensu Stace, 2019); see Section 6 for discussion.  The characteristic 'fleur de lys' fruiting catkin scales of B. pendula with reflexed lateral lobes also occur in B. pubescens highly introgressed from B. pendula, in B. celtiberica and in some B. papyrifera Marshall.
In summary (and see Section 3 re. introgression), once the pure B. pendula and pure B. nana (both diploids) are known these can almost always be fairly easily distinguished from tetraploid B. pubescens. The latter may display evidence of variable amounts of introgression from one or other of these two species. In mixed stands of B. pendula and B. pubescens the important thing is to certainly identify B. pendula (if present); all the remaining birches are likely to be tetraploid and freely inter-fertile, and hence effectively comprising a breeding unit distinct from a 'pure' B. pendula breeding population. These tetraploids can be recorded as B. pubescens.
Many of the features listed in Table 1 are illustrated below: Betula pendula (Figs. 4 -9; Betula pubescens subsp. pubescens (Figs. 10,11,21); Betula pubescens var. fragrans 22); Betula pubescens s.l. (Fig. 20).   Trees intermediate in morphology between the parents have been considered to be hybrids and reported as common and widespread (Marshall, 1914;Walters, 1975). It is suggested by Walters (1975) that their common occurrence here is correlated with the secondary status of much birch woodland in which the original ecological isolation of the two species has broken down. Further discussion of naturally occurring and artificially created hybrids is contained in Atkinson (1992) and in the Hybrid Flora (Stace et al., 2015), which reviews the pre-2000 literature.
However, while the F1 triploid hybrid (2n = 42) does occur, it appears to be rare. For example, studies in secondary mixed species birch woodland in East Anglia (Gill & Davy, 1983;Howland et al., 1995) (Nokes, 1979 (Gill & Davy, 1983;Howland et al., 1995), diploids showed less morphological and molecular variation than tetraploids. Morphologically the diploids were a good match for B. pendula, while the tetraploids were much more variable.
The range of variation in the tetraploids included some trees which corresponded well with B. pubescens and others which corresponded well with B. pendula; most of the tetraploids were intermediate in their expression of at least one, and usually more than one, of the characters measured (Gill & Davy, 1983). Forbes & Kenworthy (1973), mapped the distribution of B. pendula and B. pubescens in the whole of the upper River Dee catchment in Aberdeenshire. They commented "We have obviated the problem of separating all individuals into B. pendula and B. pubescens by taking as a unit not individual trees but stands of trees. Although every tree in a stand cannot be assigned to one species or the other, it is possible, on the basis of trees which have been positively identified, to estimate the proportion of the two species which are present in the stand. This involves the assumption that the proportion of B. pendula in the unidentified trees is the same as that in the trees which can be identified. Errors arising from this assumption are insignificant, since in practice it is seldom that more than 10 percent of the trees in a stand cannot be placed in one species or the other." This approach provides a practical solution.
However, such an approach may inflate the actual abundance of B. pendula. H. McAllister's suggestion (pers. comm.) is to certainly identify B. pendula (if present), all the remaining trees are likely to be tetraploid and freely inter-fertile, hence forming a breeding unit distinct from a 'pure' B. pendula breeding population. The tetraploids, though variable in appearance, can all be referred to B. pubescens. The above is as would be expected as diploids (2n=28) can produce unreduced gametes (n=28) which can fertilise the tetraploid (2n=56) B. pubescens, producing tetraploid hybrids inter-fertile with 'true' pure B. pubescens. There is no obvious way that tetraploid B. pubescens is likely to produce a gamete with n=14 to transfer genes from B. pubescens to the diploid B. pendula. Double reduction at meiosis is possible but likely to be extremely rare or nonexistent.
From my own observations in Strathspey, in north Scotland, it is straightforward to identify typical examples of both species, even where they occur in mixed stands, as long as determinations are based on as many identification features as possible (Table 1).

3). Introgression from the diploid (2n=28) B. pendula and B. nana, into the tetraploid (2n=56) B. pubescens.
While F1 hybrids may be rare, it has now been elegantly demonstrated (using two different molecular techniques) that introgression is widespread (Wang et al., 2014a andZohren et al., 2016). Both studies clearly separated B. pendula from B. pubescens, and both from B. nana. Using genotyping at 12 microsatellite loci, and restriction site-associated (RAD) markers, they detected introgression from the two diploid species, B. pendula and B. nana, into the tetraploid B. pubescens, with negligible indication of introgression from B. pubescens into the other two species (Zohren et al., 2016). They also found there was an introgression gradient from the north of Scotland and Orkney to the south of England. Introgression from B. pendula into B. pubescens was very low in the north, increasing to the south. Introgression from B. nana into B. pubescens was the opposite; very low in the south, higher to the north. Both clines were highly statistically significant. The evidence of introgression from B. nana is remarkable, as evidence of its admixture in B. pubescens genotypes was detected far to the south of the species' current range in Great Britain. The authors concluded that "a zone of hybridization between B. nana and B. pubescens moved northwards through the UK since the last glacial maximum, leaving behind a footprint of introgressed genes in the genome of B. pubescens".
From the supporting evidence file, Table S1, (Zohren et al., 2016), I have calculated the percentage admixture within samples of B. pubescens (Table 2). That introgression from B. pendula into B. pubescens is widespread across Great Britain, and that the degree of introgression is seven times higher in the south than in the north, is an important consideration when identifying B. pubescens. While none of these samples would be referred to B. x aurata (if that name is restricted to the F1 hybrid), it can be hypothesised that B. pubescens in the south of Britain will be more likely to show morphological features associated with B. pendula, than trees in the north of Britain. That suggestion is supported by studies in secondary birch woodland in East Anglia (Gill & Davy, 1983;Howland et al., 1995). Introgression from B. nana into B. pubescens is only occasional in the south of Britain, and the degree of introgression is merely vestigial; it is truly a 'ghost' of B. nana's former range earlier in the current post glacial period. In the north of Britain, introgression from B. nana is widespread, and while the degree of introgression is low, it is still higher than that from B. pendula. See Section 7.

4). The occurrence of distinctive local forms, especially of B. pubescens.
In their Betula monograph, Ashburner & McAllister (2013) give great emphasis to chromosome number (ploidy level) when making taxonomic decisions. Given that closely related species with the same ploidy level usually interbreed freely, within a single geographic area there is likely to be only one species with each chromosome number, unless flowering times have little or no overlap. Birches are wind pollinated, and mostly self-incompatible. Pollen can be dispersed over long distances, hence within an area there should be few differences between individuals or populations. This is a good taxonomic and evolutionary argument for lumping rather than splitting taxa, at least at species level. Nevertheless, morphologically more or less distinctive forms of B. pubescens do exist within relatively restricted geographic areas. If we were to consider the whole of Britain as a unit, we might expect all B. pubescens to be broadly similar in appearance. Even allowing for the observed gradients in introgression from B. pendula and B. nana (discussed above), B. pubescens is not homogenous.  stated that B. pubescens was (in Britain and Ireland) "A very variable species perhaps divisible into several subspecies", probably reflecting Marshall's account of Betula in The Cambridge British Flora (Marshall, 1914)  describes as "commoner among the hills of the west and north of Great Britain than in the lowlands of the south and east; on lowland peat-moors in the north of England; common on the Pennines".
Two of Marshall's taxa (B. pubescens var. microphylla E.S. Marshall and B. pubescens var. sudetica E.S. Marshall) he described as taken "together are almost sufficiently distinct from the other forms of B. pubescens to justify their being regarded as a separate species". His var. microphylla is clearly the same taxon as Ashburner & McAllister's var. fragrans, though there are discrepancies in some of the details, i.e. Marshall states the bark is brown and shining (cf. white or greywhite, sometimes brown) and the female catkins are suberect (cf. pendulous). He described it as rare in England and Wales, locally abundant in Scotland. Marshall's var. sudetica, is perhaps just a variant with narrower rhomboidal leaves, with acute apices, cuneate base and coarser teeth. He thought it rare.
Marshall described his final variety, B. pubescens var. alpigena Blytt, as a low tree or shrub, branches tortuous, leaves deeply, singly or doubly dentate, occurring between 500-700 m AOD in Scotland. It is not clear how to interpret this taxon. It is not illustrated in the Cambridge British Flora. There is material in the herbarium at Cambridge, but images are not available on line. Sell & Murrell (2018), who were based at Cambridge when writing their Flora do not refer to it, but would have been familiar with Marshall's collections.
At least two other additional forms within B. pubescens sensu lato have been described in Britain. Ashburner & McAllister (2013, pp. 313-314) mention a very distinctive form of B. pubescens (which they chose not to name) on the fringes of Dartmoor (in SW England) and in Brittany (NW France), sometimes referred to as the 'brown birch'. This is a tall tree, with whitish-brown or brown bark, velvety-hairy shoots, twigs and leaves, and twigs ± rough with brown resinous warts. It is part of the variation within their var. pubescens.
Trees closer to B. pendula in general appearance, but with persistent abaxial axillary hair tufts on leaves are tetraploids, the result of introgression from B. Betula celtiberica has been reported, mainly from Wales, as well as Arran (Chater, 2010a, Chater, 2010b, Sell & Murrell, 2018 and is included as B. pubescens subsp. celtiberica (Rothm. & Vasc.) Rivas Mart. in the 4 th edition of Stace's Flora (Stace, 2019). Sell & Murrell (2018) report it as planted in Cambridgeshire and probably planted widely elsewhere. Chater (2010b) records it as a native and occasionally planted tree in Cardiganshire (v.c.46). However, the RHS website lists no suppliers for this species, so nurseries may be listing this tree under a different name.
Tetraploid populations of Betula in an area will be expected to form an interfertile breeding unit, and so, for trees with 'celtiberica' morphology to persist, there would have to be some mechanism that created reproductive isolation, eg. period of flowering, which could be correlated with altitude. Alternatively, trees with 'celtiberica' morphology might repeatedly arise de novo through hybridisation and introgression.
These trees, very distinctive in Cardiganshire (A. Chater pers. comm.) require further study. The application of molecular techniques will be required to establish their links to other Betula populations. Currently the more parsimonious explanation is that trees with the appearance of 'celtiberica' are part of the variation within British B. pubescens, due to introgression from B. pendula, a view with which H.
A selection of images of material from Puerto de Pajares, Oviedo, Spain can be viewed at the Arboretum Wespelaar website. Fig. 23 shows typical leaves from this location. Fig. 24 shows a Cardiganshire (west Wales) example of what is identified as B. celtiberica by A. Chater and P. Sell.

5). Planting of alien (to the local area) genotypes and species, and effects of habitat disturbance.
In relatively undisturbed habitats, B. pendula and B. pubescens occupy rather different habitats, their distributions differ somewhat, and B. pubescens occurs at higher altitude. With disturbance to habitats, road and track construction, tree felling and planting, these differences have partially broken down. For example, Abernethy Forest in the Highlands is the largest of the remnant native pinewoods in Scotland. Within the core native pinewood only B. pubescens occurs. However, B. pendula is now colonising the verges of manmade tracks through the forest. On an area of bog, previously drained and planted with Pinus contorta (Lodgepole Pine), the non-native pines have been removed and the site re-wetted. The area now has dense regeneration of B. pendula, where none would have been expected before.
Provenance of planting stock is often not local. In Strathspey, in the Cairngorms, occasional B. pubescens is found as a planted tree. Such examples are clearly subsp. pubescens, whereas the usual type found in semi-natural woods locally is var. fragrans.
It is especially concerning that B. pendula is sometimes being planted beyond  The name B. pubescens subsp. pubescens has been used in all British Floras since Warburg's account . Ashburner & McAllister (2013) apply the name at variety level. Other than that the name presents no problems.
However, the small-leaved form of B. pubescens in Great Britain has gone by five different names over the last 70 years: Betula pubescens subsp. odorata (Bechst.) E.F.Warb., B. pubescens subsp. carpatica (Willd.) Ascherson & Graebner, B. pubescens subsp. tortuosa (Ledeb.) Nyman and B. odorata Bechst. All these names were incorrectly applied to this taxon. Gardiner (1984) noted that the small-leaved B. pubescens in the north west Highlands of Scotland, were distinctive cf. other populations in lowland Britain and Europe. Ashburner & McAllister (2013) agreed that these trees were distinctive. In their monograph it is named B. pubescens var. fragrans.
The incorrect or misapplied names are discussed below.

Betula pubescens subsp. odorata (Bechst.) E.F.Warb and B. odorata
Bechst. Gardiner (1984) provides a summary of Bechstein's original description of B. odorata. Bechstein described it as a single-stemmed tree, taller than B. pendula and with coarser branching, bark strikingly white, twigs pubescent without warts, leaves larger and broader, buds viscid, aromatic. Apart from the viscid buds and being aromatic, the description is almost the complete opposite of the taxon that Warburg described as B. pubescens subsp. odorata. Gardiner therefore concluded that B. odorata falls within B. pubescens subsp. pubescens. Marshall (1914) had synonymised B. odorata with his B. pubescens var. vestita, which falls within Warburg's subsp. pubescens, so it is surprising that Warburg used this name for the small-leaved taxon. Sell & Murrell (2018) elevate this taxon to a full species, B. odorata Bechst. Putting aside that it is debateable to recognise the taxon at species level, the name is incorrect. Ashburner & McAllister (2013) synonymised B. odorata with B. pubescens var. pubescens.
Betula carpatica Wild. was described from the Carpathian Mountains, and is one of several names that have been applied to disjunct populations of B. pubescens s.l. confined to mountains and hills in central and south-eastern Europe (Ashburner & McAllister, 2013). The Flora Europaea account also states that subsp. carpatica occurs in Arctic Europe, however this has been subsequently rejected (Jonsell, 2000;Väre, 2001).

Betula pubescens subsp. tortuosa (Ledeb.) Nyman
Based on a principal component analysis of leaf measurements, Gardiner (1984) synonymised B. pubescens subsp. carpatica with B. pubescens subsp. tortuosa. He noted that the name tortuosa, when used as a subspecies name, took precedence over carpatica, and hence he argued that "both the Arctic populations of Scandinavia, Finland and Russia and the montane populations of southern Europe of B. pubescens s.l. may be referred to B. pubescens subsp. tortuosa". As mentioned earlier, Gardiner also noted that the small-leaved birch in the north west Highlands of Scotland, were distinctive from other populations in Europe, and he recommended that they should be recognised at variety level (following Marshall, 1914) (Jonsell, 2000) also rejects the name B. tortuosa, agreeing with the conclusions of Orlova and Hämet-Ahti. Jonsell suggests that the 'mountain birch' is the result of introgression from B. nana, and that the correct name would be B. pubescens subsp. czerepanovii. However, because the transition to B. pubescens s.s. is completely clinal, and because the 'mountain birch' is not a homogenous taxon he does not recognise it. Väre (2001)  In Scandinavia and Iceland a more or less distinctive form of B. pubescens is dominant in the subalpine zone, often referred to as 'mountain birch (Jonsell, 2000). In recent literature (see Section 6), this is referred to B. pubescens subsp. czerepanovii, B. pubescens var. pumila, or is not given taxonomic recognition. It was formerly often incorrectly named tortuosa at species or subspecies rank. The adoption of the subspecies name tortuosa in the New Flora (Stace, 1991), wrongly indicates a close similarity of the small-leaved B. pubescens in the Highlands of Scotland, and the Scandinavian 'mountain birch'.
The northern form of B. pubescens, i.e. var. pumila, is thought to be derived from subsp./var. pubescens through adaptive selection and introgression from B. nana (Ashburner and McAllister, 2013). In Scandinavia and Iceland, both introgressed B. pubescens and its triploid hybrid with B. nana occur, the latter much more frequently than in Scotland (Jonsell, 2000, Thórsson et al., 2007. In Scotland introgression from B. nana into B. pubescens is frequent, but at a very low level (see Section 3). It is disputed to what extent such introgression is the cause of the small leaf size observed in (at least some) B. pubescens. In the first edition of the New Flora (Stace, 1991) it is suggested that "small-leaved variants of (subsp. tortuosa, sensu Stace) may have arisen by introgression from B. nana".
However, the fourth edition (Stace, 2019) says "Suggestions that the small leaves (of subsp. tortuosa) are derived from past hybridisation with B. nana are purely conjectural". The earlier suggestion was that only the small-leaved variants of the subspecies were introgressed from B. nana. H. McAllister (pers. comm. 2016) commented that var. fragrans (the widespread small-leaved type in Scotland) shows no signs of morphological influence from B. nana. He also added that he would "probably refer some Scottish Betula specimens to var. pumila, though I have not mentioned it in the monograph as occurring in the UK".
Marshall's account of Betula (Marshall, 1914) is instructive. He describes two forms of the B. nana x B. pubescens hybrid: x B. alpestris Gurke, which he described as very rare (only certainly from Ben Loyal (Fig. 27, left) and Rannoch Moor), and x B. intermedia Gurke (Fig. 27, right), which he described as rare, but recorded from several sites in Scotland. (I have retained the hybrid names used by Marshall (1914), to maintain compatibility with his account. However this naming convention is now used for intergeneric hybrids which these are not.) Marshall also commented that some forms of his var. microphylla (i.e. var fragrans) show a strong resemblance to his x B. intermedia. From Marshall's descriptions, his x B. alpestris is closer to B. nana, probably the F1 hybrid, and his x B. intermedia closer to B. pubescens. The latter appear to be backcrosses to B. pubescens (or even var. pumila, see below), being taller bushes (to 4 m) with morphology of leaves, fruits and fruiting catkin scales closer to B. pubescens (Table   3).
The Glen Callater plant (Fig. 27, right) was described as a "good sized tree" (Marshall, 1901). Another example collected by Marshall (as x B. intermedia) was from a stream side (1,400 ft.), near Bachnagairn, Clova District, v.c.90, 29/6/1904. "It was about 12 feet (3.7 m) high, conspicuously differing from the surrounding B. pubescens at a good distance by its much darker foliage, thickly interlacing branches, and peculiar rounded outline, which resembled that of a giant bush, rather than of an ordinary tree." (Druce, 1905).   Fig. 28. This tree, from the shape of the fruits and female catkin scales, is introgressed from B. nana. The leaves are very similar to the Glen Callater plant in Fig. 27, but larger, and the fruiting catkins pendulous. B. nana does not occur on Orkney, and so this tree must reflect its former presence here, or colonisation of Hoy by heavily introgressed B. pubescens at some point in the past. This example, while approaching var. pumila, is probably best recorded as B. pubescens, but a form displaying morphological signs of introgression from B. nana.  Rare, more frequent than x B. alpestris.
Not confirmed.