Mat-forming graminoid forb growing in often extensive stands due to horizontal rhizome with whitish branches, typically with rhizome branch lengths of (3)5–8(10) cm between aerial shoots. Aerial shoots ascending from rhizome, at base without or sometimes with a very few prophylls (reduced leaves without or with a short blade). Base of culms surrounded by a narrow sheath of withered leaves. Culms 10–20(30) cm, erect, stout (0.8–1.5 mm broad above the flag leaf, 1.5–3 mm below the flag leaf, smooth. Culms and leaves often strongly tinged with purple or violet.

Leaves flat or folded (convolute), with a marked mid vein forming a weak keel and numerous lateral veins slightly raised on the upper surface, smooth. Basal leaves 3–10(12) cm long, narrow (1–2 mm), tapering near apex. Culm leaves 1–3, similar to basal leaves but distinctly broader, 2–3(4) mm, flag leaf blade 2–4 cm long, attached at or below the middle of well-developed culms. Ligula short, 1(2) mm, truncate, fringed.

The units of the inflorescence of Poaceae are the spikelets, nearly always numerous in a panicle or spike-like inflorescence. Spikelets are composed of 2 glumes (bracts for the spikelet) and one or more flowers (the term used below) or rather floral units often named ‘florets’ because we do not know what is the exact flower. A flower or floret is composed of a lemma with 1 mid vein (probably the floral bract), a palea with 2 mid veins (either 2 fused bracteoles or perhaps 2 fused perianth leaves), 3 small organs called ‘lodiculae’ and essential in the opening of the flower at anthesis (possibly transformed perianth leaves or transformed stamens), 3 stamens (mostly), and a gynoecium of 2 fused carpels with 2 feathery stigmas and one seed.

Inflorescence a narrowly elongate panicle, 3–8(10) cm long, with erect branches (rarely spreading). Panicle occupying ca. 1/3 of culm length, with 6–10 nodes, the lowermost node at a considerable distance (33–40 % of entire panicle length) from the next node, interrupting the panicle. Branches 3–5 at each of lower nodes, smooth, the longest 20–40 mm, each with 1–3 spikelets in their distal part. Spikelets 5–7 × 2–3 mm, usually with 1–4 flowers (but only 1–2 functional). Bracts (glumes and lemmas) with rounded backs, a distinct mid vein and 2 indistinct lateral veins. Both glumes of nearly the same length, 5–7 mm, extending to the top of the spikelets and usually beyond the lemmas. Glumes ovate to lanceolate, acuminate and often with an extended, cusp-like apex turning inwards in the spikelet, glabrous, smooth and shiny, violet or purple in the lower and central parts but with a narrow or more often broad, golden yellow, hyaline margin and apical part. The mid vein, which often extends into the cusp, is black or nearly black and contrasts with the violet and golden parts of the glume. The coloration of the glumes (and lemmas if visible) and the mid vein are good characters for recognition of the species. Lemmas 3–5 mm, lanceolate, acute or obtuse (and often fringed), with the same coloration (margin violet and hyaline golden, narrow or broad) and marked mid vein as glumes, sparsely to moderately pilose. Palea with broad hyaline margin, smooth or slightly scabrous on keels, shorter than lemmas. Anthers violet, narrow, 2–3(3.5) mm. In most populations anthers are exerted and well-developed with full pollen grains; in some populations they are shrivelled and yellow, not fully exerted, and obviously non-functional.

Fruit an achene (with one seed).

Sexual reproduction by seeds; efficient local vegetative reproduction by extensive local clonal growth that may result in (physiologically) new individuals when the mat is fragmented by frost soil movement (cryoturbation), erosion, flooding or other accidents. Wind pollination; probably an outcrossing species.

Fruits are spread by water and birds. We also assume that birds may spread shoot fragments over shorter distances.

Dupontia fisheri s. lat. is distinguished from Arctophila by much narrower leaves, those of vegetative shoots not conspicuously distichous (with leaves in two dense rows), in glumes and lemmas much more elongated into acute, acuminate or fringed apices, and in glumes as long as the spikelet and longer than the lemmas. For differences towards Arctodupontia scleroclada, see that species.

The octoploid ('fisheri') morph of Dupontia fisheri is distinguished from the tetraploid morph ('psilosantha') by several characters: The octoploid morph has a stouter culm, often stout up to the panicle, a narrowly elliptic panicle with erect branches, and grows in mires and moss tundra rather than swamps; the tetraploid morph has a slender culm, a broad, pyramidal panicle with patent to retrorse branches, and it grows in very wet or brackish, swampy marshes.

The octoploid ('fisheri') morph of Dupontia fisheri is a typical inhabitant of sediment plains and shallow mires on fine-grained substrates, demanding permanent moisture or seepage but usually not growing submerged. It is indifferent as to soil reaction (pH) but always grows in places where there is an abundance of soil water from rivers, snow or melting permafrost. The mineral nutrient level in the stands of this plant is probably adequate.

Probably present in all zones and sections. Probably frequent (more or less) over the major islands of the Spitsbergen group but absent from Bjørnøya and not recorded from Hopen or Prins Karls Forland.

Both morphs of D. fisheri are rather common in most arctic regions but rare in Greenland (Bay 1992; Elven et al. 2011). They are also absent from the N Russian regions west of the White Sea and from NW Europe, however, present in Svalbard. The connection is therefore probably to the east, to NE European Russia.

Dupontia has by some authors been considered a monotypic genus, i.e., with only one species: D. fisheri. Other authors have accepted, at least three species have been described: D. fisheri s. str., D. pelligera and D. psilosantha (Tzvelev 1976; Czerepanov 1995). The genus is one of the very few almost exclusively arctic genera of vascular plants but is certainly hybridogeneous. Brysting et al. (2003, 2004) showed that Arctophila fulva is part of the parentage and that the entire genome of Dupontia has been homogenized in the direction of Arctophila. They checked several other alternative genera looking for the second parent but did not find any plausible candidates. From the current genetic evidence, Arctophila and Dupontia could be merged as one genus with Dupontia as the priority name (Robert Brown in 1823, in Chloris Melvilliana). However, the morphology is rather different and until recently authors have upheld Dupontia as a genus separate from Arctophila (e.g., Cayouette & Darbyshire 2007b). In a recent study (Tkach et al. 2020), however, Arctophila is merged within Dupontia as D. fulva Röser & Tkach. If this approach is followed, also Arctodupontia with A. scleroclada must be included, as Dupontia scleroclada (Rupr.) Rupr.

There are some peculiarities with respect to the chromosome numbers and ploidy levels in Dupontia. The following main numbers are reported (Elven et al. 2011): 2n = 42–44 (hexaploid if x = 7, tetraploid if x = 11), 2n = 84–88 (dodecaploid if x = 7, octoploid if 2n = 11), and 2n = 126–132 (18-ploid if x = 7, dodecaploid if x = 11). It might indicate that the second parent has a basic chromosome number different from x = 7 but that some stabilization in the direction of the normal grass number (of x = 7) occurs. Several reports have been made of numbers intermediate between these three levels, also from Svalbard (Brysting et al. 2004; Brysting & Elven 2005). However, the three ploidy levels, whatever their background is, are represented in the field by three rather different plants. We here denote them as tetra-, octo- and dodecaploid morphs. The tetraploid morph has slender culms, very open panicles with spreading to patent (or even retrorse) branches, and is confined to very wet marshes (fresh or brackish) or shallow waters, often together with Arctophila. The octoploid morph has culms stout up to the panicle, narrow panicles with mostly erect branches, and occurs in shallow marshes and mires, much less inundated than the tetraploids. The dodecaploid morph resembles the tetraploid morph in panicle (very open with patent to down-pointing branches) but is very coarse with culms stout up to the panicle. The dodecaploid morph is known from marshes (fresh and brackish) in N Canada, Beringia and Siberia, but not from Svalbard.

The three morphological appearances associated with the ploidy levels have been shown, by molecular means, to have arisen repeatedly and independently in the three investigated regions: in Svalbard (only two of the levels), Canada and Russia (Brysting et al. 2004). This means that the tetraploids of Svalbard, even if morphologically similar to the tetraploids of Canada and Russia, are genetically more closely connected to the octoploids in Svalbard (and similarly in other areas). We are not aware of any similar observations among arctic plant species. The conclusion is that the morphologically and ecologically rather clearly different entities in each region cannot be treated as different taxa because morphologically inseparable plants in other regions have a different origin. An extensive summary of this discussion (with many references not cited here) is found in Elven et al. (2011). Cayouette & Darbyshire (2007b) consequently accepted only a single species of Dupontia without races for the Flora of North America.

The question one has to face, as field botanists in Svalbard, is how to treat an evident and ecologically important variation that does not conform to the (convincing) genetic data and does not fit into the ordinary taxonomic framework. The solution we have chosen is to describe the two morphological, cytological and phytogeographical entities we see as 'morphs', i.e., informal morphological expressions of genetic differences. We accept that many populations are intermediate in morphological expressions, possibly also in chromosome numbers (and the percentage of stands with aborting anthers in Svalbard is fairly high). Already Scholander (1934) described the variation, including the intermediates.

Bay, C. 1992. A phytogeographical study of the vascular plants of northern Greenland – north of 74 northern latitude. – Meddelelser om Grønland, Bioscience 36. 102 pp.

Brysting, A.K., Aiken, S.G., Lefkovitch, L.P. & Boles, R.L. 2003. Dupontia (Poaceae) in North America. – Canadian Journal of Botany 81: 769–779.

Brysting, A.K. & Elven, R. 2005. Tundragras Dupontia fisheri og hybriden med hengegras Arctophila fulva rundt Ny-Ålesund, Svalbard. – Blyttia 63: 186–193.

Brysting, A.K., Fay, M.F., Leitch, I.J. & Aiken, S.G. 2004. One or more species in the arctic grass genus Dupontia? – a contribution to the Panarctic Flora project. – Taxon 53: 365–382.

Cayouette, J. & Darbyshire, S.J. 2007b. Dupontia R. Br. – In: Flora of North America Editorial Committee (eds.), Flora of North America north of Mexico. 24. Magnoliophyta: Commelinidae (in part): Poaceae, part 1: 602–604.

Czerepanov, S.K. 1995. Vascular plants of Russia and adjacent states (the former USSR). – Cambridge University Press, Cambridge.

Elven, R., Murray, D.F., Razzhivin, V. & Yurtsev, B.A. (eds.) 2011. Annotated Checklist of the Panarctic Flora (PAF) Vascular plants. http://panarcticflora.org/

Scholander, P.F. 1934. Vascular plants from northern Svalbard with remarks on the vegetation in North-East Land. – Skrifter om Svalbard og Ishavet 62. 155 pp.

Tkach, N., Schneider, J., Döring, E., Wölk, A., Hochbach, A., Nisen, J., Winterfeld, G., Meyer, S., Gabriel, J., Hoffmann, M.H. & Röser, M. 2020. Phylogenetic lineages and the role of hybridization as driving force of evolution in grass supertribe Poodae. – Taxon 69: 234–277.

Tzvelev, N.N. 1976 [1977]. Zlaki SSSR. – Nauka, Leningrad.