The viviparous plants of the Poa arctica–pratensis complex (except for P. pratensis ssp. colpodea, see separate entry) span the range of morphological variation between P. arctica ssp. arctica and P. alpigena. Even if we often find viviparous plants with the habitus of typical P. arctica ssp. arctica, or of typical P. alpigena, a large proportion of what we find is something in between. This may be due to the genetic background of the viviparous plants. They may be of hybrid origin between the two species or between incompatible strains within each species. We therefore have decided not to try to make a separation we would find artificial and rather present a common description without much measurements.

Mat-forming graminoid herb growing in often extensive stands due to extravaginal, branched rhizome, typically with rhizome branches of up to 10 cm between shoots. Aerial shoots ascending from rhizome, at base with several prophylls (reduced leaves without or with a short blade). Culms to 40 cm (occasionally taller in manured sites), erect, smooth or minutely scabrous just below the panicle. Base of aerial shoots with a few withered leaves forming a loose sheath.

Leaves keeled, flat or folded, smooth. Basal leaves narrow, tapering evenly or at the apex, much shorter than culms. Culm leaves usually 2–3, similar to basal leaves or slightly broader, flag leaf blade attached around or above middle of culm. Ligula 1–3 mm, subacute to acute. Leaves and culms glabrous. Lower parts of leaf sheaths, blades and culms green or sometimes tinged purplish, rarely tinged deep purple. Plants with red pigments usually approach P. arctica in other features.

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).

Inflorescence an open, elongate or pyramidal panicle with erect, spreading or sometimes retrorse branches; panicle occupying 1/4–1/5 of culm length. Panicle with 4–12 nodes and 2–5 branches at each of the lower nodes, smooth or very sparsely scabrous. Branches 20–30(40) mm long, the lower ones with up to 10 spikelets (fewer in plants approaching P. arctica in other features), concentrated to the ends of branches (in plants approaching P. arctica) or along larger parts of the branch length (in plants approaching P. alpigena). Spikelets usually with 3 flowers. Bracts (glumes and lemmas) with keels, often fairly sharp. Glumes narrowly or broadly lanceolate, acute, ca. 1/2 as long as spikelet or longer, with (1)3 indistinct veins, glabrous, smooth or slightly scabrous on the keel in upper (distal) parts, violet, purplish or often green at base and without or with a very narrow white hyaline margin. Lemmas lanceolate or broadly lanceolate, acute, with 5 more or less distinct veins, with more or less wavy (curly) hairs on the veins and keel but usually glabrous between the veins, with a large tuft of cottony hairs at the base of the lemma, often hyaline on the distal parts of the keel, violet, purplish or green at base, with a narrow hyaline margin variegated in bronze yellow and white. Paleas with pubescent veins and shorter than lemmas. Stamens 3 (if present); anthers shrivelled and non-functional.

The entire spikelet except the glumes (and sometimes the lowermost flower) is transformed into a bulbil (vivipary), with green or slightly violet or purplish leaves.

The florets in a spikelet transformed into a leafy bulbil.

Vegetative reproduction by bulbils. Mat-forming and possibly with some vegetative reproduction by fragmentation of stands. Bulbils germinate to 97 % in an experiment (Alsos et al. 2013).

In some grass genera, Poa among them, there is a fairly good correlation between ploidy levels and modes of reproduction. Diploids (mostly 2n = 14) and tetraploids (mostly 2n = 28) usually have sexual seed-set, whereas higher ploidy levels are often (but not always) asexual with either seed-set without fertilization (agamospermy) or vegetative propagation by bulbils replacing the flowers in the spikelet (vivipary).

The viviparous plants in the Poa arctica/pratensis group all have efficient local growth by the rhizomes and may form extensive stands that can be fragmented by soil movement (cryoturbation, solifluction), other disturbance, or simply through age. Still, the only mode of efficient dispersal is the bulbil. The bulbils are green (or green colour covered by violet or purplish pigments) and are virtually vigorous small plants, not yet rooted in the soil. They are shed regularly at the end of the season and probably constitute a more significant part of the total reproduction in the viviparous types than seeds do in the comparable seminiferous types. For some favourable adaptive traits in bulbils, see Poa alpina var. vivipara.

Bulbils are easily spread by wind and partly with animals. Some dispersal of rhizome fragments is possible.

The viviparous plants often assigned of Poa arctica and P. alpigena are readily distinguished from the seminiferous plants of the same species by producing bulbils instead of fruits, and the entire group of P. arctica–pratensis (seminiferous and viviparous alike) differs from all other Svalbard Poa by branching outside the leaf sheaths (extravaginal), producing horizontal rhizome branches of various length.

Within the Poa arctica/pratensis group, the viviparous plants have been attempted divided on 3–4 entities, all identified in significant parts of the herbarium material: P. arctica var. vivipara, P. pratensis ssp. alpigena var. vivipara, P. pratensis ssp. colpodea, and P. herjedalica (assumed to be the hybrid or a hybrid species from P. alpina × P. pratensis). Except for P. colpodea, which is rather distinct, we have not been successful in dividing the viviparous group. Poa colpodea seems to be a taxon in its own right (as also concluded by Haugen 2000) and is given a separate description. There are, however, few characters separating P. colpodea from the other viviparous parts of the group. We here refer the differential characters emphasized in the key of Haugen (2000) for P. colpodea: In P. colpodea the culms are ca. 9–15 cm, the panicle short, 2.4–3.2 cm, the panicle branches short, 1.3–7.0 mm, sometimes scabrous, not spreading, the lower glume usually 3-veined, 2.6–3.5 mm, about as long as upper glume, the keel of the upper glume smooth or scabrous, the lemma ca. 0.68–0.85 mm wide, never hairy between veins, with slight tomentum at base, and anthers usually absent; in the other viviparous forms the culms are taller, 12–20(50) cm, the panicle longer, 3.3–5.2 cm, the panicle branches long, 6.1–14.1 mm, smooth, often spreading, the lower glume often 1-veined, 3.3–3.8 mm, 80–90 % as long as upper glume, the keel of the upper glume usually smooth, the lemma ca. 0.86–0.97 mm wide, glabrous or sparsely hairy between veins, with well developed webby tomentum at base, and anthers usually present, but shrivelled.

The most evident difference in the field is that P. colpodea has a constantly very narrow panicle with very numerous small spikelets along much of the branches, all developing into bulbils and into a dense, bushy inflorescence (or whatever one should call a collection of bulbils). The other viviparous plants of the P. arctica–pratensis complex have a more open panicle, often pyramidal, with larger spikelets concentrated at the ends of the branches, and never attain the bushy appearance of the ‘colpodea’ panicle.

Viviparous Poa arctica/pratensis is as common and widespread as its seminiferous counterparts and occurs in both stable and unstable vegetation covers, in meadows, heaths, moist tundra, bird cliff vegetation, and also on very open ground, sometimes as the sole vascular plant species (e.g., in boulder screes). It is also largely indifferent as to soil moisture, ranging from the tussock level in mires and moderately soaked sediment plains to dry ridges and screes, but with its main occurrences in moist to moderately dry sites. Some of the viviparous plants are very strong competitors and take over the dominance, especially in fertilized sites, e.g., bird cliff meadows. They are largely indifferent as to soil reaction (pH), present on both acidic and basic substrates.

Viviparous Poa arctica/pratensis types are largely independent of climate and found in a broad range of site types and probably throughout the islands, i.e., along the entire gradients from middle arctic tundra to polar desert zones and from weakly oceanic to clearly continental sections. They are found more or less throughout the investigated parts of Svalbard (including the outlying islands of Kong Karls Land and Hopen). There are very few records from Edgeøya and Barentsøya but we suspect this to be due to under-collecting. Scholander (1934) noted that he had only a single collection from Nordaustlandet (from the most favourable site on the entire island: Floraberget at Murchisonfjorden in Gustav V Land); all the other viviparous plants belonged to P. colpodea. His comment should perhaps warn us that the records from the even more harsh places like Kong Karls Land and Hopen should be reconsidered; they could belong to P. colpodea.

Outside Svalbard, viviparous forms of Poa arctica and/or the P. pratensis group seem to be rarities even if some bulbil-reproducing species or races are described as closely related to one or the other. In Scandinavia, there may be some single bulbil-reproducing plants here and there, but only P. lindebergii Tzvelev in the Dovrefjell mountains in S Norway is a regular, viviparous species or race with a range. Its affinity is not clear, but has been suggested as developed from a hybrid P. arctica x P. flexuosa (see Elven et al. 2022). In Greenland and arctic Canada, the only bulbil-reproducing Poa is P. colpodea. The situation in the Russian arctic is less clear, but we suspect that the pattern in Svalbard may be repeated in some of the islands in European Russia and perhaps east to Taimyr, probably not farther east. In E Siberia, the bulbil-reproducing Poa is, again, P. colpodea.

Seminiferous plants are much more easily assigned to either P. arctica or species of the P. pratensis group than are viviparous plants. The most obvious reason for this is that some of the most reliable characters characterizing the two species are found in the lemmas, and the lemmas are often totally transformed into minute green leaves without any diagnostic features whatsoever in the bulbils. This development starts so early that it is often not possible to find a spikelet with intact lemmas. Another reason is more hypothetical. Vivipary in grasses is often assumed to be a reproductive escape from sterility in hybrids. The viviparous strains of the P. arctica–pratensis complex may accordingly belong to hybrids between the two species or between incompatible strains within the complexes (we suspect that both these possibilities are realized). In viviparous plants, the only moderately efficient features for separation are found in panicle shape, number of branches per node in the panicle, number of spikelets per branch and their position along the branch, and whether branches are scabrous or not. We have not been able to divide the material into smaller groups according to such characters. It is a common experience to find viviparous plants closely resembling either P. arctica or the P. pratensis group in general appearance; however, when one looks at the details, one often find combinations of characters from the two species or species groups. Haugen (2000) collected what she assumed to be viviparous P. arctica ssp. arctica but concluded, after analysing the material morphologically, that it had to be assigned to P. alpigena due to the lack of hairs or spinules between the veins on the lemmas (the single most effective diagnostic character between P. arctica s. lat. and P. pratensis s. lat.). Her conclusion may have been premature (see below).

Another problematic plant is P. herjedalica. This is assumed to be a hybridogeneous species from the cross P. alpina × alpigena (ssp. alpigena). Plants assigned to P. herjedalica on the mainland are usually identifiable by having ‘socks’ of sheaths at the base of the shoots and large spikelets (from P. alpina) combined with some extravaginal shoots breaking through the ‘socks’ (from P. alpigena). Poa herjedalica is rather frequent in the Scandinavian mountains (e.g., Hylander 1953), at least as numerous separate hybrid clones, but much less stabilized than the hybrid species P. jemtlandica (from P. alpina × P. flexuosa; see Brysting et al. 1997, 2000). We consider P. herjedalica to be a swarm of primary hybrids reproducing by bulbils rather than a taxon in its own right. Specimens annotated as P. herjedalica constitute a not insignificant part of the Svalbard material of Poa in the herbaria, especially among the often very scanty and imperfectly conserved older collections. The presence of this hybrid in Svalbard, and in rather large amounts, is highly improbable as one of its assumed parents, P. alpina, is extremely rare there as native and seminiferous (i.e., regularly producing pollen and stigmas). There is no correlation between the localities reported for P. herjedalica and the few known sites of native, seminiferous P. alpina. We therefore, for the time being, consider all reports of P. herjedalica from Svalbard as erroneous and as belonging to some part of the viviparous P. arctica–pratensis complex.

This is the part of Poa where the study of Haugen (2000) is of little help. Among the rhizomatous plants, she accepted two seminiferous taxa (as P. arctica ssp. arctica and ssp. caespitans) and two viviparous ones (as P. alpigena ssp. colpodea and var. vivipara). Seminiferous P. alpigena was excluded from her sampling as that taxon was outside the proposed and probable parentage of the two assumed hybrid taxa she studied (P. arctica ssp. caespitans and P. hartzii). Haugen (2000) found, however, enzymatic support for a separation of P. colpodea from her viviparous P. alpigena var. vivipara. In her UPGMA analysis of the isoenzyme data (Haugen 2000: 72), the majority of her P. alpigena var. vivipara mixed with P. arctica ssp. arctica. We now assume that the viviparous forms bridge the gap between seminiferous P. arctica ssp. arctica and P. alpigena but that the two species are rather well separated.

Alsos, I.G., Müller, E. & Eidesen, P.B. 2013. Germinating seeds or bulbils in 87 of 113 tested Arctic species indicate potential for ex situ seed bank storage. – Polar Biology 36: 819–830. Doi 10.1007/s00300-013-1307-7.

Brysting, A.K., Elven, R. & Nordal, I. 1997. The hypothesis of hybrid origin of Poa jemtlandica supported by morphometric and isoenzyme data. – Nordic Journal of Botany 17: 199–214.

Brysting, A.K., Holst-Jensen, A. & Leitch, I. 2000. Genomic origin and organization of the hybrid Poa jemtlandica (Poaceae) verified by genomic in situ hybridization and chloroplast DNA sequences. – Annals of Botany (London) 85: 439–445.

Elven, R. 1974. Artsinnvandring og vegetasjonsutvikling på resente morener i Finseområdet. – Cand. real. Thesis, Univ. Oslo, Oslo.

Elven, R., Bjorå, C.S., Fremstad, E., Hegre, H. & Solstad, H. 2022. Norsk flora. Ed. 8. Det Norske Samlaget, Oslo.

Haugen, J. 2000. Possible hybrid origins of Poa hartzii and Poa arctica ssp. caespitans (Poaceae) investigated by morphology and isoenzymes. – Cand. scient. Thesis, Univ. Oslo, Oslo.

Hylander, N. 1953b. Nordisk Kärlväxtflora, I. – Almqvist & Wiksell, Stockholm.

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.