3  Arctic vascular plants

3.1  Natives and introductions

Some scientists consider the whole Arctic as strongly influenced by human activities like all other parts of the globe. Others consider it as a predominantly virgin environment. Both views are right, depending on points of view. As to the phytogeography of vascular plants, the latter viewpoint is closest to the truth, as indicated by the following three features:
            (A) As far as we know, no single described, predominantly arctic vascular plant species has yet gone extinct due to human activities in the last 250 years. This is not only true for the Arctic in general but also probably for each of the six countries involved.
            (B) Neither is the Arctic strongly influenced by 'aliens', i.e., plants with their origin elsewhere but spread to the Arctic through human activities. The total number of aliens established in the Arctic (and fully included in the Checklist) is low, and their proportional part of the flora is very low to absent in all regions. The introduced plants are – with few exceptions – confined to settlements and their close surroundings, trading posts, mining areas, airstrips, harbours, and the few roads and railways. They are not very expansive and they are not considered as a danger to the native flora. The exceptions are found in the North Atlantic regions where present and/or past permanent and extensive agricultural settlement for a thousand years or more has made more impact, that is, in SW Greenland, Iceland and arctic N Norway with settlements back to Viking age (in N Norway perhaps even much earlier), and in some Russian areas with more recent settlement. In these areas agricultural archaeophytes constitute a significant part of the now long established flora. The time frame in the Arctic for these originally introduced plants is now more than a millennium, i.e., probably the same as or longer than for some natives. The lack of aggressive immigrants in the Arctic is naturally the result of the possible aggressors being adapted to much more ameliorate conditions, climatically and otherwise.
            (C) The native historical plant geography is intact in the Arctic. There are extremely few, if any, documented case of spread of plants by man among arctic regions. In many temperate areas such spread is disturbing and obscuring the 'natural' patterns – patterns caused by non-anthropogenic physical and evolutionary processes – to a degree that makes a separation between non-anthropogenic and anthropogenic plant geography practically impossible. There has been frequent travel between some arctic regions, especially between W Alaska and E Chukotka, between N Norway, Svalbard, NE Greenland, and Iceland, and between SW Greenland and NE Canada, but we know of no fully substantiated case of arctic trans-regional anthropogenic dispersal of native plants. Some cases have been proposed from Greenland, but the possible source areas of these plants are not arctic.
            Together these three features make the Arctic phytogeographically different from the other major biomes in the World (except for the Antarctic and some high-alpine areas). It makes the Arctic into a still virtually virgin field for studies of the influence of non-anthropogenic processes on phytogeographical patterns, e.g., the influences of the Pleistocene glaciations and the lines back to the early Quaternary and Tertiary past. This will, however, probably not last. Recent changes strongly influencing the Arctic include anthropogenic climatic change, 'modernization' of the communities of its native peoples, and increased use of the Arctic, both for recreation, for research, and for development of its mineral resources, will by necessity strongly influence its flora and phytogeographical patterns. The Checklist may therefore be a reference point before this process accelerates.
            The records in the Checklist are therefore separated on three categories based on age in the Arctic (whether native / archaeophytic or more recently anthropogenic) and permanence:
            (1) Records as native / archaeophytic. These include plants we assume having reached the Arctic by non-anthropogenic means, and plants brought to the Arctic by the early (at least pre A.D. 1700) activities of native people or early agricultural immigrants. For these plants all occurrences we know of are included in the Checklist, also stray records of plants that have appeared in the Arctic by 'natural' means (transported by, e.g. birds or sea currents). These 'native casuals' are too few to necessitate separate consideration: Suadea (probably) in Russian Far East, Cakile in Svalbard and perhaps in Greenland, Atriplex in Bear Island, Gnaphalium in W Greenland, and perhaps a very few others.
            (2) Records as stable introductions. These include plants obviously brought to the Arctic by man's activities later than approx. A.D. 1700 and self-sustaining in their arctic sites by seed reproduction for at least more than one generation, or by vegetative means. The principle has only been possible to follow in the comparatively few places in the Arctic where there is permanent settlement and where botanists go fairly often, mainly in Nome and some other settlements in W and N Alaska, Churchill and a few other places in N Canada, and numerous places in Greenland, Svalbard, Iceland, and mainland Norway. For the Russian parts it is mainly guesswork, but guesswork based on what we know generally about the ecological and reproductive requirements of the relevant taxa. We have included rather than excluded dubious cases. This means that perennial taxa with some means of vegetative persistence in the places where they have been introduced ('remaining' taxa) usually have been included, whereas annuals and biennials with comparatively high thermal requirements for reproduction and perennials without vegetative spread have been excluded. The stable introduced plants are – almost without exception – confined to past or present settlements or trading posts. We know of no single case where they have had an extensive spread from their assumed places of introduction, except for some recent and probably not fully stable spread on road verges and along railways.
            (3) Records as casual introductions. Casuals are quite frequent in some arctic settlements, sometimes very far to the north. Due to the relative density of settlements, and the frequent communication with Denmark, a very high number of such exotics are recorded for W–SW Greenland, much higher than in, e.g., N Iceland or mainland N Norway. The reason is probably that the Greenland communities have been and are much more dependent on import than the more self-sustaining communities in Iceland and Norway. Some other hot-spots of exotics are also fairly well documented, especially Nome in W Alaska, Churchill (railroad endpoint) at Hudson Bay, the Vorkuta area (railroad endpoint) in NE European Russia, in Siberia the Dudinka area at lower Jenisei River and Tiksi close to the Lena River delta, and in Russian Far East especially the townships of Anadyr and Provideniya. A similar situation to that in Greenland is evident in the large industrial settlements in arctic Russia, especially Vorkuta, Norilsk – Dudinka, Tiksi, and Provideniya, nearly totally dependent on imports for their existence. Even in the four high arctic Russian and Norwegian mining settlements in Svalbard (at 78–79°N) are records of, e.g., sunflowers, peas, apples and pears, and of a few for these latitudes quite exotic weeds (e.g., Atriplex sagittata and Medicago polymorpha) scarcely recorded as stable even in temperate N Europe. For the main parts of the Checklist, casual records are only included for plants that are documented or assumed as stable in at least one arctic region. We have made no full survey of other such casual records, but those we have come across are listed in Appendix 1.

3.2  Borderline arctic occurrences

The southern boundary of the Arctic is 'fuzzy', both in the nature and in our treatment. As the treeline is a very important ecological boundary, numerous species reach or more often slightly cross this line. Even more diffuse is the arctic boundary in the treeless coastal regions bordering on the North Pacific and North Atlantic. We have been fairly liberal in inclusion of plants that only occur at the boundary as drawn for the Checklist, also to be prepared for future mapping of changes, e.g., northwards spread of (comparatively) thermophilous plants. Plants that occur at or very close to the southern boundary of the Arctic as we have drawn it are therefore consistently included but always indicated as 'borderline' (Part A Taxonomy) or marked as 'b' (Part B Distribution).
            The number and 'quality' of borderline records that we have accepted vary between regions. In Greenland there are none, as all of Greenland is considered as arctic. In Iceland, mainland N Norway, and the NW European Russian Murman area, the arctic boundary is fairly exactly drawn, and the plant distributions well recorded (for Iceland Kristinsson's maps (http://vefsja.ni.is/website/plontuvefsja), for N Fennoscandia including the Russian Murman area, e.g., Hultén 1971, for the Murman area in more detail Pojarkova 1953–1966). In these regions the indicated borderline occurrences are quite real.
            Further east in Russia the boundary is less well defined and drawn and also the floristic mapping of records less exact. There are also many discrepancies among different sources, e.g., among Flora Arctica URSS, Flora Sibiri, and the Russian Far East Flora. The main initial Russian input to the Checklist has been from the scientists of the Komarov Botanical Institute in St. Petersburg, but very often different data and different interpretations have been presented by the Siberian research groups. Resolving such Russian controversies must be the responsibility of the Russian scientific community.
            Within Russia we have adhered strictly to the arctic boundary as proposed by Yurtsev. We have therefore included as borderline cases many species with documented occurrences at or north of that boundary even when Russian botanists have stated that these species are 'not arctic'. We agree with their interpretation of most of these species, but as long as they occur at or north of the boundary as proposed by Russians, we can not exclude them. This implies an uncommonly large number of borderline cases along some of the major Russian rivers, especially Pechora, Ob, Jenisei, Lena, Kolyma, and partly Anadyr. In most or all of these cases, the arctic boundary should probably have been drawn further north along the rivers than in the surroundings (as it is along, e.g., Mackenzie River in NW Canada). We have observed that the true (and sharp) Larix treeline along Lena River lies c. 100 km north of the boundary as drawn in the Russian proposal.
            In North America, the boundary follows the treeline except for the oceanic regions of SW Alaska and NE Labrador, but the number of borderline cases is still probably too high in some regions due to records being less exactly mapped. The boundary southwest of Kuskokwim Mts in SW Alaska is especially arbitrary and also important because this is a species-rich area with very numerous northern extensions of Pacific species. The number of borderline cases would be high wherever we drew the line. Similarly, the inclusion of all of Brooks Range, Philip Smith Mts, British Mts, and northernmost Richardson Mts north of the treeline results in quite numerous slightly transgressing boreal and Cordilleran plants as borderline cases. The lower Mackenzie River poses the same problem as the large Siberian rivers, but here our boundary is drawn at the northern treeline. Still, the Mackenzie delta is a hotspot of borderline arctic taxa. The arctic boundary in non-forested NE Labrador are drawn straight to the sea from the inland treeline and is therefore arbitrary. Still, there are much less borderline cases here than in other coastal areas, mostly due to the species-poor flora.
            The absolutely highest number and relative amount (percentage) of borderline cases for any region in the Checklist are recorded at S Hudson Bay, c. 65 taxa. The real number may be even higher as taxa with clear occurrence at the coastline of the bay are entered by us as 'rare' and not as 'borderline'. Exclusion of the S Hudson Bay coastline from the Arctic as circumscribed would probably reduce the arctic flora as treated by the Checklist by 100–150 taxa. At the same time, this is an area with extensive marshes and shores dominated by otherwise arctic species (and a heavy concentration of Polar Bears) and is seen as truly arctic by most visitors. We have to accept that this is a region where the Boreal and the Arctic overlap more extensively than perhaps anywhere else in the Arctic.

3.3  Taxonomic categories

The Checklist consistently applies the following categories: Family, genus, species, and subspecies. Other categories are included more sporadically and only when based on some criteria: infrageneric categories (i.e., categories between genus and species), varieties, and hybrids.
            Families. Families are treated very conservatively. There are numerous recent studies on molecular markers and cladistics that throw doubts on many of these traditional families, e.g., Scrophulariaceae (Olmstead et al. 2001) and Primulaceae (Källersjö et al. 2000). There will be many dramatic changes in families if we continue from the Checklist phase to production of an up-to-date Panarctic Flora.
            Genera. We have recognized genera more narrowly than in many current regional treatises. Here we have often taken into account recent molecular and phylogenetic studies, but not to the degree currently done even in systematic textbooks. The genera will probably also change appreciably if a Panarctic Flora is realized.
            Infrageneric categories. Such categories are only included in the Checklist in a few comparatively large genera (at least > 10 species) and only where they can be applied more or less unambiguously. This is in contrast with many Russian treatments where a major focus in later years has been on infrageneric classification. Our approach is more in accordance with, e.g., the Flora Europaea tradition where these categories were applied most sparingly. It is often problematic to decide where an infrageneric classification is relevant, and there are also many cases in the arctic flora where there seems to be even more disputes about these categories than about genera and species. One explanation may be the arctic speciation processes. If allopolyploidal speciation is prevalent among arctic vascular plants, as many scientists assume, the higher ploidal levels may be based on crosses between diploids in different sections and series. In such cases a subgeneric classification may be relevant for the diploids but less so for the often more numerous polyploids, due to reticulations. There are strong indications of such reticulations in many large arctic genera like Draba, Oxytropis, Poa, Potentilla, and Salix. Instead we often apply the informal category 'aggregate' where there is a need for a common comment on a species group.
            Species (excluding agamospecies). As to species (and also subspecies / varieties), there are many different traditions in how to handle these categories. The problems concerning arctic taxa are discussed in several contributions in Nordal & Razzhivin (1999).
            In general, we probably circumscribe species more broadly in the Checklist than done in most previous (and some current) Russian treatments but more narrowly than in some current North American treatments. Species is applied (generally) for taxa more or less disjunct in several assumed independently inherited characters and with mostly assumed reproductive barriers towards related species. This must be based on assumptions as very little factual evidence is available for most arctic plants, even if this situation is rapidly improving. One important feature is that we, in principle, consider documented or assumed allopolyploids as species in their own right, and not as hybrids, when they have an independent reproduction and an independent range (i.e., more than a 'first' hybridization event or a hybrid swarm).
            Geographical criteria have often been applied previously in many Russian treatments for distinguishing closely related species and subspecies. Largely sympatric entities (with strongly overlapping ranges) and fully allopatric entities (disjunct or vicariant) have in these tretaments mostly been accepted as species, whereas parapatric entities (with partly overlapping ranges and often with intermediates in the zone of overlap) have been considered as subspecies. These restrictions are not applied for this Checklist. We accept as species entities in all the three categories of geographical pattern. Proven or assumed reproductive isolation is applied as a species criterion where appropriate (and where known).
            However, there are several problems with the reproduction criterion. Firstly, many arctic plants are inbreeders, and it is more or less irrelevant or unknown whether they can or really do cross regularly. Secondly, there are indications that reproductive barriers may arise more rapidly when and where entities are in contact than where they are not. From a 'biological species' viewpoint, the former should be species whereas the latter could be subspecies. Close to nothing is known about the potential or realized reproductive behaviour of arctic plants in this aspect. For all practical purposes we have to rely on morphology and guesswork. A major part of the discussions among the PAF collaborators have concerned these matters. In most cases, the Russian participants have preferred treatment as species, the North American participants have preferred treatment as subspecies or just a collective species, and the NW Europeans have vacillated (e.g., in several groups of grasses, the Anemone narcissifolia group, the Astragalus australis group, several groups in Oxytropis, some Antennarias, and the Arnica angustifolia group). But, there have also been several cases where NW Europeans and North Americans go for species and Russians for subspecies or a collective treatment (e.g., Huperzia, the Epilobium davuricum – arcticum group, and the Pedicularis sudetica group). Thirdly, there are major groups of arctic plants that reproduce asexually (apomictically).
            Agamospecies. Asexual (apomictic) reproduction is assumed mostly to be results of initial hybridization and subsequent escape from hybrid sterility by asexual means. Large groups of arctic plants propagate mainly by asexual means. For these groups almost all the reproductive criteria for species recognition are irrelevant. These groups also differ among themselves in most relevant aspects: Morphological variation patterns, degree of apomixis (whether facultative or nearly obligate), degree of polyploidization, means of propagation, and the geographical patterns. The cases span from: (a) Some occasional change to (apomictic) bulbil reproduction as in a few species of Bistorta, Deschampsia, Festuca, and Saxifraga; (b) Occasional agamospermy in a single or a few species in a genus as in Draba and possibly in Alopecurus; (c) Complexes of sexual functional diploids, primary hybrids, and agamospermic functional polyploids ('functional' because the basic chromosome level in the genus may be, e.g., tetraploidal or higher) as in Antennaria, Arnica, Calamagrostis, Erigeron, Pilosella, Poa, and Potentilla, in Poa also complicated by bulbil-reproducing entities; to: (d) Full-fledged large and mainly polyploidal agamospermic complexes with sometimes hundreds or thousands of described 'species' as in Alchemilla, Hieracium s. str., the Ranunculus auricomus group, and Taraxacum. The amounts of such apomictic taxa are perhaps highest in areas strongly affected by the Pleistocene glaciations in temperate parts of Europe (there also with, e.g., Rubus and Sorbus) and North America (there also with, e.g., Amelanchier, Crataegus, and Symphyotrichum). This concentration is often interpreted as resulting from the disturbances and migrations over the entire continents during the glaciations. The percentage of apomicts among described taxa is, however, also high throughout the Arctic.
            The apomictic complexes have been treated in many ways throughout their systematic history. One approach has been to try to identify morphologically the smallest entities and name them, mostly as species. This has resulted in description in the Nordic area alone of c. 2000 species of Hieracium, c. 400 species of Taraxacum, and c. 400 species of the Ranunculus auricomus group, and from the rest of Europe several hundreds or thousands of species of the same genera and of Alchemilla and Rubus, among others. Other parts of the world have not had the same eager 'species-namers', but some North Americans working on Antennaria approach the Europeans in this respect. This approach is clearly non-functional from the viewpoints of biodiversity, conservation, and practicality. If these entities are considered as 'full' species, the absolute majority of named and geographically restricted, vulnerable or threathened 'species' on the European continent would be agamic microspecies of 5 or 6 genera, separable only by very minute criteria, and recognizable only by a few specialists, some of them now long deceased. Another approach was exemplified by Flora Europaea (Tutin et al. 1964–1980). In this Flora, the authors of these genera included the numerous microspecies in aggregates informally named 'groups' or 'circle species', where a more widespread microspecies was defined as the descriptor of a group of more restricted microspecies. This approach has also been followed in Russian treatments of, e.g., Hieracium and Taraxacum. A third approach is exemplified by the North American treatments of Antennaria by Bayer and Chmielewski, recognizing a few widely circumscribed species with subspecies, by Bayer also with separation of the assumed basal sexual entities from the assumed derived agamic entities as subspecies.
            The situations in the different apomictic groups are probably too divergent to allow for a common approach. For the Checklist we have applied several approaches:
            (a) In the genera where there are just a few agamic entities in the Arctic, separable from their assumed sexual progenitors (extant or perhaps extinct) by sufficient morphological criteria, they have been treated just as normal species. Examples: Bistorta, Deschampsia, Erigeron, Saxifraga, Solidago, and Symphyotrichum.
            (b) In the genera where there are at least some extant basal functional diploids and derived agamic polyploids of known or assumed parentage, they have partly been treated as species, partly as subspecies, depending on their morphological distinctness. Examples: Calamagrostis, Festuca, Pilosella, and also Antennaria which is a borderline case towards the next category due to its high number of described agamospecies.
            (c) In the genera or groups consisting predominantly or entirely of agamic entities, and where very numerous agamospecies have been described, the Flora Europaea approach has been applied, grouping the named arctic microspecies into fewer 'collective species' comparable in morphological width and geographical range with sexual species. Examples: Hieracium, the Ranunculus auricomus group, and Taraxacum. One exception here is Alchemilla, where each of the accepted arctic representatives is quite distinct and entered in the Checklist as such, even if this genus is a mess further south, e.g., with more than 300 named European agamospecies.
            (d) For two partly apomictic genera the PAF collaborators have not fully agreed on an approach: Poa and Potentilla. In both cases the discussion is mainly about the importance of hybridization in the origin of named entities, about whether the evolutionary process may be mapped with any certainty without genetic investigations (or even with genetic investigations), and whether it is worthwhile or futile to describe, name, and accept numerous assumedly hybridogeneous agamic lineages as taxonomic entities. The situation is perhaps that the analysis of the morphological variation in these genera – and speculations about its causes – has arisen at a much later stage than in some other agamic groups like Alchemilla, Hieracium, Rubus, and Taraxacum. We have at least to accept the historical importance of the 19th and early 20th century systematists describing the perhaps 5–6000 species in these genera; we are more reluctant to do the same today. In parts of Poa the discussion concerns whether three major mixed sexual and agamospermic groups, partly also with reproduction by bulbils, are to be treated more collectively (as proposed by, e.g., Soreng) or with numerous species (as proposed by Tzvelev): the P. arctica group, the P. pratensis group, and the Stenopoa – Tichopoa section(s). Parts of the Checklist entry for this genus is more or less disapproved on by all PAF participants. In Potentilla, the situation is more complicated as there is some experimental support for the hybridization – agamospermy model underlying the treatment of Soják and Yurtsev resulting in very numerous 'species', but there is a reluctance on the part of at least Elven, Eriksen and Murray to accept all the assumed hybridogenous products as taxa without more support. Also for this genus, most PAF collaborators disagree with a smaller of larger part of the Checklist entry.
            Subspecies. The application of the ranks of subspecies and variety is at present quite different among the majority of non-Russian European, Russian, and North American botanists.
            Russian botanists have applied (and probably still apply) the rank of subspecies for parapatric, assumed interfertile entities whereas they (as stated above) prefer to apply the rank of species for both sympatric and allopatric situations. This is a mainly geographically based criterion. They apply the rank of variety for more local, ecotypical, or even modificational variation.
            The majority of North American botanists currently engaged in the Flora of North America project apply the ranks of subspecies and variety as interchangeable. They apply the rank first given to a taxon. This is also supported in part by the current International Code of Botanical Nomenclature (ICBN Vienna Code, McNeill et al. 2006) where entities named at these two ranks are treated as homonyms. In the treatments published until now in Flora of North America, the application of rank (as subspecies and/or variety) depends on the preference of the author.
            Non-Russian European botanists treat the two ranks of subspecies and variety as different, as exemplified by Flora Europaea (1964–1980, 1993), Flora Nordica (2000 onwards), and almost all regional floras (e.g., Elven in Lid & Lid 2005 for Norway). This is also the treatment applied for this Checklist. The category 'subspecies' is applied for morphologically definable major geographical or eco-geographical races and sometimes for ecological races and ploidal races (assumed to have arisen through autopolyploidy) if they also can be morphologically characterized. If different genomes are assumed to be involved (allopolyploidy), the normal category for the Checklist is as species. This is a pragmatic and debatable definition as it does not include any criterion concerning reproductive barriers. Reproduction data, as stated above, are largely absent for arctic plants and are also probably irrelevant for allopatric entities. In practical cases, geographically and ecologically overlapping entities that keep distinct are treated as species whereas intergrading entities are treated (mostly) as subspecies, in the latter case depending on their morphological distinctness. See, e.g., the Checklist discussion concerning Salix bebbiana. For geographically non-overlapping entities, and for geographically overlapping but ecologically non-overlapping entities, the criteria are purely morphological. If the entities are morphologically close, they are mostly treated as subspecies. Assumedly non-adaptive features are given more weight than assumedly adaptive ones in deciding between rank as species or subspecies (or variety, see below).
            Varieties. The category 'variety' is only applied very sparingly in the Checklist. The absolute majority of named varieties have either been omitted or, mostly in the case of North American 'varieties', rather been accepted at rank of subspecies if relevant combinations are published. Except for these cases where available combinations restrict us, we apply the category variety mainly for ecological races, ecotypes, or for very local but morphologically distinct entities (see, e.g., Artemisia globularia var. / subsp. lutea). Only exceptionally well defined and widely recurrent varieties are included in the list. 'Variety' is also used for some local (geographical) races where the morphological and assumed genetical differentiation is slight. Inclusion of varieties in the Checklist is accompanied by notes.
            The chosen treatment of the two categories – subspecies and variety – was decided initially in the PAF Checklist work (see Elven 1999): "Both species, subspecies and variety shall be viewed pluralistically, meaning that several different biological situations might be covered by each, based on different criteria. They are to be viewed as levels of differentiation only." and "Species, subspecies and varieties shall be based on morphological criteria. Available data on genetics, evolution, and reproductive system should be taken fully into account, but should not overrule morphological evidence." No evolutionary connection is assumed, i.e., we do not assume that variety is a developmental stage towards subspecies, or subspecies towards species. In some cases it may be so, in others not.
            Hybrids and hybridogeneous taxa. Primary hybrids and hybrid swarms, even if named and accepted in regional Floras, are – as a rule – not accepted as taxa for the Checklist. Hybrids are only included where they have an internal homogeneity and an existence independent of their proposed parents, either by occurrence well outside the range of one or both parents or by independent propagation (sexual or asexual), or preferrably both. See, e.g., discussions in connection with Coptidium x spitsbergense, Poa x jemtlandica, xPucciphippsia vacillans, and Trollius x apertus. Inclusions of such hybrids are argued for by a note.
            Hybridogeneous taxa, i.e., entities originating from some past hybridization process but reproductively isolated from their parents, mostly by ploidal difference (e.g., Anthoxanthum odoratum, Poa annua and Saxifraga nathorstii) are treated as full species and not as hybrids, not least in view of the prevalent polyploidy in the arctic flora. Most arctic species have probably arisen in this way. The criterion is whether they have been stabilized and have attained a range of their own. There are obvious borderline cases, especially in Potentilla, and also the complications of agamic reproduction, this also especially in Potentilla. Due to the vast differences among the cases, and among the opinions of the PAF collaborators, there is little consistency in this aspect in the Checklist. It may (and should) improve in a future Panarctic Flora.