It is worth checking old data – validation of Asplenium onopteris L . presence in the most northeastern sites in Europe ( Sudetes , SW Poland )

In Poland, isolated serpentine rocks are exclusive habitats of some Asplenium species, reaching here their north or northeastern border range. One of them was Asplenium onopteris, a diploid European species native to Mediterranean and Atlantic areas. Since the nineteenth century, Polish out-of-range sites of A. onopteris have been quoted in literature without critical verification. Thus, to verify occurrence of this species in Poland, we analyzed the nuclear DNA content and micromorphological features as well as critically reviewed the literature data. We proved that all individuals from Polish populations resembling A. onopteris were tetraploids and should be classified as A. adiantum-nigrum. In addition, we validated a taxon silesiacum reported as co-occurring with A. onopteris. The proposed diagnostic features are insufficient to indisputably delimit this taxon, and distinguishing it as a separate unit is not justified. Analyses of the DNA content revealed also the presence of a triploid A. ×centovallense, a new hybrid for Polish flora.


Introduction
The European Asplenium adiantum-nigrum complex includes three related and morphologically variable taxa: diploid A. cuneifolium Viv. and A. onopteris L. and their stable tetraploid hybrid A. adiantum-nigrum L. [1].Asplenium cuneifolium is associated with serpentine rocks in Southern Europe, up to southern Germany and southwestern Poland [2].The most frost-resistant, circumpolar, and cosmopolitan A. adiantum-nigrum can colonize also other types of rock [3]; its European range covers the Atlantic and sub-Mediterranean area up to western Norway, southwestern Poland, Slovakia, Ukraine, and Romania [4].Asplenium onopteris is the most thermophilous species, tolerant to the substratum type, with a continuous range in Macaronesia, northwestern Africa, and the Mediterranean area [4,5].The species has also been confirmed in a discontinuous location in Ireland [5,6] and reported from Poland [6,7].In Poland, all species of the complex with the only Central European population of A. onopteris and the locus classicus of A. silesiacum Milde are associated exclusively with serpentine rock outcrops in the Sudetes.The latter taxon is currently included in A. adiantum-nigrum though its origin, taxonomic rank, and diagnostic features vary and are sometimes contradictory [8][9][10].Both last mentioned taxa were of particular interest for botanists due to suggested endemism and the distance to the continuous range.
Our research primarily focused on biology and ecology of peripheral populations of the complex due to their biological importance (e.g., [11][12][13]), in order to estimate the habitat conditions limiting the A. onopteris occurrence to only four out of several serpentine rock outcrops in Poland.The long-term continuous monitoring of A. onopteris populations revealed enormous morphological variability related to the fern age and habitat conditions.Therefore, the main goal of the presented research was to validate the identity of the ferns quoted in the literature as the only Central European populations of A. onopteris.Also, we aimed to elucidate the current status of the cooccurring taxon "silesiacum".We applied three approaches: (i) a critical review of the published data; (ii) estimation of the nuclear DNA content by flow cytometry, and (iii) morphological analyses.

Material and methods
To check the changes in the nomenclature of the revised taxa, we have reviewed all available publications since the first paper by Milde [14].
The field research was conducted between 1996 and 2016 on all populations of A. onopteris and A. silesiacum reported in the Sudetes (SW Poland).As comparative material, we used A. adiantum-nigrum and A. cuneifolium sampled in the neighboring serpentine rock outcrops and A. onopteris collected in its continuous distribution range, in Pitigliano (Tuscany, Italy), Bari Sardo (Sardinia, Italy), and Tenerife (Canary Islands, Spain).
To determine the species identity, we estimated the nuclear DNA content and measured the size of the spores and stomata.Due to the rarity of the A. adiantum-nigrum complex in Poland, only one leaf was taken from each tuft.Leaves were collected from all ferns resembling A. onopteris and A. silesiacum sensu Milde (41 samples).As a reference, 31 leaves of A. onopteris, 31 of A. cuneifolium, and 34 of A. adiantum-nigrum were analyzed.
The nuclear DNA content was analyzed for all taxa studied from the Sudetes populations (18 leaves of potential A. onopteris plants, eight of A. cuneifolium, and 10 of A. adiantum-nigrum), and additionally for A. onopteris from Pitigliano (three leaves).Samples were prepared according to Galbraith et al. [15], with some modifications.Prior to the genome size analysis, the presence of propidium iodide (PI) staining inhibitors in plant tissues was tested [16].Leaves of Vicia faba 'Inovec' (2C = 26.90pg) [17] were used as an internal standard.Plant tissues of the target taxa and internal standard were chopped in 1 mL of Tris-MgCl 2 nucleus isolation buffer supplemented with 2% (w/v) polyvinylpyrrolidone (PVP-10), PI (50 μg/mL), and ribonuclease A (50 μg/mL).For each sample, 7000-10 000 nuclei were measured using a Partec CyFlow SL Green flow cytometer (Partec GmbH, Germany).At least five replicates were performed for each taxon.The obtained histograms were analyzed using FloMax software (Partec GmbH, Germany).The genome size was estimated using the linear relationship between the ratio of the target taxon and internal standard 2C peak positions on the histograms.The mean coefficients of variation of the 2C DNA content were estimated for all samples.The 1C genome size was calculated after the conversion of values into base pair numbers using the formula 1 pg = 978 Mbp [18].
The size of stomata and spores was analyzed in voucher materials.Prior to the measurements, basal fragments of dried pinnae were soaked with warm water for ca. 1 h and the strips of abaxial epidermis were peeled.Specimens were analyzed and documented by an Olympus Microscope BX50 -DP71 camera -Cell^B software system (Olympus Optical, Poland).The length of the spores (with perispore) and stomata was measured with AxioVision LE 4.2 software (Carl Zeiss MicroImaging GmbH, Germany).For each taxon, six to eight specimens were analyzed.For each specimen, 100 randomly selected stomata and 30 spores, when available, were measured.The results were analyzed in Statistica 12.
At that time, Lower Silesia was a province of Prussia but A. onopteris was not included in the flora of Germany [26,27] until 1903, when it was mentioned as A. adiantum-nigrum subsp.onopteris Heufl.with var.silesiacum Milde [28].In the same year, the shortened name written as "A.onopteris Heufl." was used by Schube [29] in reference to the Gozdnik population as a separate name of the species and a synonym of A. adiantum-nigrum subsp.onopteris (L.) Heufl.The author reported also its new stands in the Kiełczyńskie Hills (Ślęża Massif) and near the village of Kamionki (Sowie Mts), both with f. silesiacum.Pax [30] quoted all data and presented A. onopteris f. silesiacum as an endemic taxon to the Ślęża Massif serpentine rock.In the second edition of Hegi's Flora, the taxon was named by Bergdolt A. onopteris var.silesiacum ( [31] and re-editions in 1961, 1965).In the latest edition, A. silesiacum Milde was synonymized with A. cuneifolium and described as a form similar to A. onopteris [32].It was also classified as A. cuneifolium var.silesiacum Dostál [33].
Field research on A. onopteris focused mostly on the Gozdnik Hill population.The presence of A. onopteris was noted here, with var.silesiacum [48] or subsp.silesiacum [49] as an endemic taxon.Żołnierz [50] questioned the distinctiveness of "silesiacum" as a separate taxon but reported the presence of A. onopteris and added its new stand from the Kamienny Grzbiet Hill (Ślęża Massif); the chromosome number was cited after the literature, without revision.Next, A. onopteris and f. silesiacum were recorded in the Gozdnik Hill but big and glossy Asplenium ferns in the Kiełczyńskie Hills were classified as A. cuneifolium Viv.var.anthriscifolium Milde [51].On the other hand, the presence of the species was also questioned in Kamionki (after [52]), and in general in Lower Silesia [53].

Nuclear DNA content
Three different values of the 2C DNA amount, corresponding to three ploidy levels, were obtained in the studied plants.The determined 1C genome size of all specimens ranged from 4690 Mbp to 8993 Mbp (Tab.S1).The 2C nuclear DNA content for the diploid A. cuneifolium was 9.59 pg, and similar for the diploid A. onopteris from Italy (9.04 pg), whereas for the tetraploid A. adiantum-nigrum, the 2C DNA amount reached 18.45 pg (Fig. S1a-c, Tab.S1).In the ferns from the putative population of A. onopteris in the Sudetes, the genome size was 18.38 pg/2C and typical of A. adiantum-nigrum tetraploid plants (Fig. S1d, Tab.S1).The presence of the diploid A. onopteris was excluded at all stands -almost all examined plants similar to A. onopteris were tetraploids.The only exception were some ferns collected in Kamionki.Their nuclear DNA content was of about 13.90 pg/2C, approximately in between diploid and tetraploid species.These specimens represented A. ×centovallense D. E. Meyer (Fig. S1e, Tab.S1), a triploid hybrid between A. cuneifolium and A. adiantum-nigrum; both parental species co-occurred at the stand.In the Kiełczyńskie Hills, only A. cuneifolium was found.
The obtained histograms of the 2C DNA content displayed distinct G 0 /G 1 peaks (Fig. S1) with a mean coefficient of variation (CV) ranging from 4.82% for the analyzed diploid plants, 4.83% for triploids, to 4.87% for tetraploid plants.

Size of stomata and spores
The stomata size varied mostly in A. adiantum-nigrum and in the representatives of the putative A. onopteris from the Sudetes.In the remaining ferns, stomata size variability was less pronounced and similar in analyzed populations (Fig. 1, Tab.S1).The smallest stomata, with mean lengths ca.46 µm, were present in A. cuneifolium, whereas for other taxa, mean values were between 52 µm and 60 µm.In addition, the values in diploid A. onopteris overlapped with those in A. adiantum-nigrum.Stomata in ferns from the Sudetes resembling A. onopteris, which according to our DNA analysis were tetraploids, were of a similar size as in A. adiantum-nigrum and had ca.60 µm in length.Ferns identified as a triploid hybrid, A. ×centovallense, had stomata ca.54 µm long.Spore sizes were more related to the ploidy level than stomata sizes and differed between diploid species (A.onopteris and A. cuneifolium), reaching 36-38 µm, and tetraploid A. adiantum-nigrum, for which spores had ca.42 µm.Spores in ferns resembling A. onopteris were of similar length to the latter species, having ca.45 µm (Fig. 1, Tab.S1).Sporangia in the hybrid A. ×centovallense were not properly developed and did not produce spores.

Discussion
The history of out-of-range Asplenium onopteris and endemic A. silesiacum taxa in the Sudetes started in the nineteenth century, when the peculiar fern population in the Gozdnik Hill was described by Milde [14,19].At that time, identifications of Asplenium taxa were based only on macromorphology.The enormous phenotypic variability of the polyploid complex A. cuneifolium -A.onopteris -A.adiantum-nigrum [1] often resulted in erroneous diagnoses.Isolated populations of the allotetraploid hybrid A. adiantum-nigrum can be indistinguishable in gross morphology from the diploid A. onopteris and were sometimes deposited in herbaria under this name [5,54].Therefore, to indisputably identify the Asplenium taxa, we analyzed the amount of nuclear DNA.
These analyses revealed three ploidy levels: diploid A. cuneifolium and A. onopteris from Italy, tetraploid A. adiantum-nigrum and revised populations from the Sudetes, and a triploid hybrid A. ×centovallense.The genome size of A. onopteris and A. adiantum-nigrum established in this study was slightly larger [55] and of A. cuneifolium slightly lower [56] than previously reported, however, these differences could result from the different buffers and standards used.The identification of a triploid hybrid is the first record of this particular taxon for the Polish flora.It is classified as a European nothotaxon A. ×centovallense D. E. Meyer nothosubsp.centovallense = A. cuneifolium subsp.cuneifolium × A. adiantum-nigrum subsp.adiantum-nigrum [54,57] (Fig. S2).We report a 2C DNA C value for this triploid hybrid, to the best of our knowledge, for the first time.It is similar to that given for the triploid A. ruta-muraria L. [55] and displayed in the Plant DNA C-value Database [58].Regarding the DNA content, all studied Asplenium species would be defined as a group with intermediate genomes (1C > 3.5 pg < 14.0 pg) [59].
In polyploid series, micromorphological features are often used as indirect markers of the taxon, related to its ploidy level [60][61][62][63][64][65].However, in the A. adiantum-nigrum complex, we found the stomata size less useful in taxon delimitation, varying in populations of a given taxon with values overlapping and similar to each other, regardless of the taxon ploidy.A comparable variability was earlier reported in A. onopteris and A. balearicum Shivas [66].Likely, it is a genus-dependent feature, in some ferns not related to the ploidy level [60,62,63], and invalid in the A. adiantum-nigrum complex.More stable feature was the spore size, slightly differing between diploid (A.onopteris and A. cuneifolium) and tetraploid species (A.adiantum-nigrum); importantly, in the latter taxon, probably due to its hybrid origin, values overlap with those of both parental species.Numerous abortive or undeveloped sporangia in Asplenium specimens indirectly indicated the presence of a triploid hybrid [57], similarly to other fern complexes.
Three vouchers of the typus of A. silesiacum collected by Milde survived in the Herbarium in Berlin-Dahlem (B; Germany).They were revised and classified as A. adiantum-nigrum subsp.adiantum-nigrum var.silesiacum [54,67] (Fig. 2).This name was proposed exclusively for the serpentine ecotype similar to A. cuneifolium in gross morphology, in contrast to var.adiantum-nigrum, which avoids this kind of substratum [8,9,[68][69][70].This proposition has not been maintained, nor is valid for Polish populations of Asplenium adiantum-nigrum as all of them occur on serpentine rocks and are clearly distinguishable from A. cuneifolium.Moreover, all examined ferns of the A. cuneifolium-shape were diploids.The other proposition was to distinguish the A. adiantum-nigrum subspecies based on the origin of maternal genes and highlighting their morphological differences [10]: silesiacum attributed to the maternal genes of A. cuneifolium, and adiantum-nigrum to those of A. onopteris.Following this classification, the population in the Gozdnik Hill (locus classicus of A. silesiacum) currently presents only features of subsp.adiantum-nigrum.Young ferns or specimens growing in open sun have transiently features of subsp.silesiacum, which disappear in older or growing in shade tufts.These findings indicate that the taxonomic position and distinguishing features of silesiacum are not fully recognized yet and suggest that it is only a synonym of A. adiantum-nigrum.
A disjunctive distribution is quite a common phenomenon in ferns [71][72][73].Homosporous ferns disperse via wind-blown haploid spores, potentially transported over thousands of kilometers.Due to the ability to reproduce through intragametophytic selfing, new fern populations can develop from a single gametophyte, i.e., from a single spore [74,75].In this light, the presence of the disjunctive population of A. onopteris distant to ca. 500-600 km to the species continuous distribution area cannot be ruled out.However, more suitable habitats can be found between its continuous range and the population analyzed but, remarkably, there is no data on its presence in well-recognized floras of Germany and the Czech Republic [76,77].The only German population of A. onopteris, documented in 1914, was disproved as containing exclusively tetraploid ferns and finally classified as A. adiantum-nigrum [78].
To conclude, we verified the identity of supposed A. onopteris and proved that the examined populations from the Sudetes should be classified as A. adiantum-nigrum; thus we disprove the presence of the species in Poland.

Fig. 1
Fig. 1 Comparison of the stomata (a) and spore size (b) in studied taxa of the Asplenium adiantum-nigrum complex.