Panarctic Flora

630508b Astragalus robbinsii var. harringtonii (Rydb.) Barneby

Geography: American Beringian: (ALA).

Notes: Murray and Elven: This section, the Astragalus australis aggregate, is one case among many of groups with a disjunct range composed of geographically isolated and/or overlapping parts throughout the mountains and tundras of southern and central Europe, Asia, and North America. The majority of these groups are absent from or present only in marginal parts of the main regions covered by the Pleistocene glaciations in northern Europe, Greenland, and northeastern North America, suggesting a Quaternary split of a late Tertiary or early Pleistocene species. The mostly allopatric to parapatric, morphologically separable parts of such groups have been treated differently among cases and among authors.

The majority of North American authors from Barneby (1964) to Cody (1996), Welsh (2007), and Gillett et al. (2007) lump all variation within one species, Astragalus australis (L.) Lam., Fl. Franç. 2: 637 (1779) [Phaca australis L., Mant. Pl.: 103 (1767); lectotype: Tilli, Cat. Pisani 19.5.14.f.1, illustration of "Astragaloides Alpina supina glabra, foliis auctioribus", 1723 (Podlech in Turland and Jarvis 1997: 479)]. Astragalus australis was described from central Europe and is reported from there with 2n = 32 (4x, at least four reports) and 48 (6x, at least five reports). The American authors argue that the morphological variation is too complex to justify acceptance of several distinct species. Note the not very elegant treatment of Scoggan (1978c) where he sorted the American variation into varieties 'named' as a, aa, b, bb etc. based on (a) leaflet shape, (b) upper leaf surface hairiness, (c) hair shape, and (d) fruit hairiness. He identified plants fitting 15 of the 16 theoretically possible boxes in this scheme. He did not take the ploidy variation into account.

Hultén (1968a, 1968b) accepted two species: the North American A. aboriginorum and the Eurasian A. australis. This solution is countered by close morphological connections across the Bering Strait, and the ploidy variation was neglected also in his treatment.

Yurtsev (1986, PAF proposal) accepted several species of which four reach the Arctic: A. gorodkovii Jurtz. 1968, A. kolymensis Jurtz. 1968, A. tolmaczevii Jurtz. 1986, and A. tugarinovii Basil. 1924. For the names based on North American plants, he had the following solution. He accepted A. aboriginorum Richardson 1823 and considered the name A. richardsonii E. Sheld. 1894 a synonym for it, but he did not accept this species to reach the Arctic. He rather included the arctic North American plants that had been assigned to A. richardsonii within A. tolmaczevii, and he considered the North American name A. lepagei Hultén 1950 a synonym for the earlier described Russian A. tugarinovii.

As to chromosome numbers, Löve and Löve (1975a) sorted, after their manner, the species that reach the Arctic according to ploidy levels: two diploids (A. gorodkovii and A. lepagei), one tetraploid (A. richardsonii), two hexaploids (A. australis and A. tugarinovii), and one octoploid (A. kolymensis). This is not fully supported by the evidence. An additional, northern but non-arctic A. australis subsp. glabriuscula from Quebec is reported to be diploid (Gervais et al. 1999, three counts).

Our provisional solution is based on a survey of northwestern North American and some northeastern Asian material (ALA, O) and on field experience with both American and Asian plants. We see a geographically structured morphological variation that does not fit with a solution with only one collective species. The morphological differences are supported by differences in ploidy levels (2x-12x) even if two or three levels are found in several of the taxa we accept. We share Yurtsev's view that there are several recognizable taxa and also his view that they are best considered at rank of species. Transitions have not been proved. We do, however, not find support for Yurtsev's proposal to fit all the arctic North American material into the species described from Asia. The older North American names A. aboriginorum and A. richardsonii cannot without further study be assumed to represent non-arctic plants only and also the name A. lepagei should be re-considered. Our more specific proposals are as follows.

(1) All the northern Asian and North American plants differ morphologically from the European mountain plant A. australis s. str. They should be recognized as species apart from the European plant.

(2) The morphological variation pattern in northern Asia and northwestern North America is geographically structured with entities with consistent, albeit often strongly overlapping, ranges. There is correlation between morphology and ploidy levels. Several taxa are definable and the large degree of parapatry and the ploidy differences do not support a model of intergrading races.

(3) The variation in arctic and near-arctic northwestern North America can be divided, more or less, on three taxa (plants): one (a) in the boreal-alpine and southern arctic parts of Alaska and the Yukon Territory, one (b) in arctic northeastern Alaska and northwestern Canada, and one (c) fairly narrowly amphi-Beringian.

Plant (a) distinctly resembles the Cordilleran plant named as A. aboriginorum. We should therefore not exclude this old name from our considerations without further study.

Plant (b) was synonymized by Yurtsev with the Siberian A. tolmaczevii (described from Taimyr) by extending its range across northern Yakutia and Chukotka to northwestern Canada. We have compared plants from northern Yakutia and Wrangel Island with American plants and are not convinced that they belong in the same species. We rather consider the arctic northeastern Alaskan and northwestern Canadian plants a separate species for which the name A. richardsonii might be the most relevant one.

Plant (c) occurs on both sides of the Bering Strait. Hultén (1950) named the Alaskan plants A. lepagei. It differs fairly distinctly morphologically from the two other northwestern North American species. Yurtsev assigned the plant in the Chukchi Peninsula to A. tugarinovii (described from Taimyr) but it differs from that species as it appears in Yakutia and West and South Chukotka (the Anadyr area) and is morphologically inseparable from the western Alaskan one. This was also noted by Yurtsev in comments. We therefore accept A. lepagei and transfer the Chukchi Peninsula plant to it.

As to the northeastern Asian plants, A. tolmaczevii differs from A. kolymensis and A. tugarinovii and probably merits rank and name. Astragalus tugarinovii and A. kolymensis are not very different but we have not seen enough material to propose a merger.

We enter the A. australis group as an aggregate of species as proposed by Yurtsev but with large modifications on the North American side not approved by Yurtsev. The collective is entered with seven species with arctic occurrences. We distinctly prefer specific rank for taxa differing in morphology and ploidy level, in addition to (overlapping) geographical ranges.

A summary of the ploidy levels (perhaps incomplete): diploid - A. aboriginorum (Cordilleran), A. gorodkovii (the northern Urals), and A. "australis" subsp. glabriusculus (non-arctic northeastern North American); diploid and tetraploid - A. tolmaczevii (northern Asian) and A. lepagei (amphi-Beringian); diploid, tetraploid, and hexaploid - A. tugarinovii (northern Asian); tetraploid and hexaploid - A. australis s. str. (non-arctic European); hexaploid - A. richardsonii (American Beringian); and octoploid and dodecaploid - A. kolymensis (northeastern Asian).