Panarctic Flora

3416 Deschampsia P. Beauv.

• P. Beauv., Ess. Agrostogr.: 91, 160 (1812).
• Casual: Deschampsia danthonioides, D. elongata.
• Excluded: Deschampsia glauca (see several notes below), D. mackenzieana.

Notes: Aiken, Elven, Murray, and Tzvelev: We consider the genus Deschampsia in a restricted meaning and keep Avenella and Vahlodea apart as genera. A molecular study based on nuclear and plastid sequences (Chiapella 2007) gives support to Avenella and Vahlodea as genera apart from Deschampsia. In addition to morphological features, Avenella and Vahlodea have the base chromosome number x = 7 and include mostly regularly diploid or tetraploid species (2n = 14, 28), whereas Deschampsia s. str. has the base chromosome number x = 13 and may be result of an ancient hybridization between different genera.

The northern representatives of Deschampsia are closely related and have all sometimes been considered races of one species: D. cespitosa (e.g., Tzvelev 1976). The taxonomy of the group is unresolved in spite of numerous regional studies (e.g., Lawrence 1945; Kawano 1963, 1966; Tzvelev 1976; Chiapella 2000; Chiapella and Probatova 2003). The differential characters are few and many features are plastic, at least in general habit (all plants are tussocky) and leaf shape and size. Some useful characters are found on the leaf surfaces. Panicles are generally of the same shape but vary in, e.g., density and in scabrosity of branches. Spikelets are predominantly two-flowered and essentially similar with silvery hyaline glumes, lemmas, and paleas, but with some useful variation in general and relative size of parts, color, callus hair length, and awn length and position of attachment.

We assume that several taxa should be accepted as species and also that there at present is no very satisfying way of doing so. No efficient comparison has yet been undertaken between Eurasian and North American plants. Some taxa are therefore circumscribed in different ways, especially D. cespitosa itself. The current western European and Russian view is that D. cespitosa s. str. is restricted as native to Europe and western Asia but widespread elsewhere as an adventive. Other authors consider it boreal to arctic circumpolar as native, and as polymorphic (previous North American opinion, including Hultén's, and partly also the current one, see Chiapella et al. in Soreng et al. 2003 and Barkworth 2007b). Another much disputed plant is D. glauca, considered by Scandinavian authors a regional Scandinavian subspecies of D. cespitosa confined to river and lake shores (e.g., Elven 1994; Mossberg et al. 2003; Elven et al. 2005), by Russian and North American authors as nearly circumpolar and not nearly as ecologically specialized. Also other species have been considered differently in different regions, e.g., D. sukatschewii (including "borealis"). This means that application of species names has been fairly arbitrary.

The two main ploidy levels are functionally diploid (2n = 26) and tetraploid (2n = 52) but there are numerous reports of intermediate numbers, perhaps due to hybridization and reticulation among species and/or among ploidy levels within species. Assignment of chromosome number reports to taxa is often made difficult, or impossible, due to the unresolved taxonomy and identification problems. The high base number of x = 13 suggests an old tetraploid with aneuploidy from 2n = 28 to 26.

We consider the traditionally wide concepts of both D. cespitosa s. str. and D. glauca unsatisfactory and assume that several native taxa must be recognized in northern Asia and in North America, perhaps also in northern Europe, all of them different from the two mentioned European taxa. Tzvelev (1976) treated the northern Eurasian taxa as numerous subspecies of D. cespitosa s. lat. Now he prefers them as species (PAF proposal). Chiapella et al. in Soreng et al. (2003) treated the boreal and arctic North American taxa, except for D. brevifolia (but that only due to lack of a valid combination), as four subspecies of D. cespitosa and two additional varieties. The differences in principle between these two approaches are not large. However, treatment as an extensive series of races does not simplify anything, especially not nomenclature. There is little or no hard evidence for intergradation of races but that mainly because the races are not strictly circumscribed and characterized. Still, we find little support for intergrading subspecies (more than twenty accepted by Tzvelev 1976). Also problematic is the lack of correspondence in names applied between northern Asia and North America and between Europe and North America. As far as we know, there is no investigation combining evidence from morphology, ploidy levels, and molecules. The study of Chiapella (2007) included only a few accessions of this group from northwestern Europe.

We have chosen an approach close to Tzvelev's PAF proposals. The variation is not a case of vicariant allopatric or parapatric races. Co-occurrence of two or more morphologically separable species within a region is the rule rather than the exception: in northern Europe D. cespitosa s. str. and D. alpina; in the High Arctic D. sukatschewii and D. brevifolia, in the North Atlantic areas also D. alpina; in northern Siberia D. anadyrensis, D. brevifolia, D. obensis, D. sukatschewii, and D. vodopjanoviae; and on the Bering Sea coasts D. beringensis, D. sukatschewii, and at least two other taxa. We assume more similarity among arctic plants in Asia, America, and Europe than done by North American authors but we do not share the wide concept of D. cespitosa subsp. cespitosa as applied by Chiapella et al. in Soreng et al. (2003) and by Barkworth (2007b). We suspect that one or two of the taxa and names still are misinterpreted. For these reasons, we mostly have accepted the taxa as species, only two with races.