Flora of North America species comparison

Species ca. 20 (8 in the flora).

Section Pensylvanicae, as here presented, accommodates most perennial species with pinnate to subpinnate leaves and styles ca. 1 mm that are frequently papillate-swollen proximally. Plants are often notably glandular, leaflet margins are usually revolute, and petals are most commonly about the same length as the sepals and often somewhat erect at anthesis. Section Multifidae was commonly used for this group of species until J. Soják (1987b) noted that sect. Pensylvanicae was an older, validly published name.

About half of the North American members of sect. Pensylvanicae have been previously included in a broadly defined Potentilla pensylvanica, with or without infraspecific divisions (for example, C. L. Hitchcock and A. Cronquist 1961b; S. L. Welsh 1974; R. L. McGregor 1986b; H. A. Gleason and Cronquist 1991). In support of a more narrowly circumscribed P. pensylvanica, B. L. Kohli and J. G. Packer (1976) found that major morphologic types coincide with three different ploidy levels. Most plants with fully pinnate leaves (five to nine leaflets in remote pairs) and other characters of typical P. pensylvanica in the narrow sense are tetraploid; plants with subpinnate leaves (mostly five leaflets in approximate pairs) are octoploid (or sometimes tetraploid; Packer, pers. comm.). Diploid populations from the northern Great Plains, morphologically closest to P. pensylvanica in the narrow sense but with epicalyx bractlets usually longer than the sepals, were named P. finitima Kohli & Packer (a later synonym of P. lasiodonta).

The ploidy differences noted by B. L. Kohli and J. G. Packer (1976), in combination with morphology (vestiture, glandularity, leaf dissection, and features of the epicalyx) and ecogeographic correlates, serve collectively to delimit the species that are recognized here. The most distinctive vestiture type is a velvety covering of short (0.2 mm), straight, spreading hairs that generally characterizes Potentilla lasiodonta and P. pensylvanica in the narrow sense, although this feature is sometimes obscured by longer ascending hairs or is missing altogether. Leaves of P. lasiodonta and P. pensylvanica in the narrow sense are also distinctly pinnate, with the blade comprising at least half the total leaf length and being no more than about half as wide as long, usually bearing at least three pairs of lateral leaflets gradually reduced in size from the distalmost, and the leaflets themselves being pinnately toothed with 4–12 blunt teeth per side. Epicalyx bractlets in both species, and also in P. litoralis, tend to be nearly as long as the sepals (to twice as long in P. lasiodonta) and concave, at least in fruit, with revolute margins and prominent midveins. These diagnostic features are not consistently present and/or evident, especially in depauperate and immature specimens, but collectively they serve to identify most specimens with a reasonable level of confidence. Occasional hybrids are to be expected.

In addition to the species addressed here, Potentilla lyngei Jurtzev & Soják subsp. lyngei has been reported from northeastern Greenland. At least one purported voucher (annotated by Jurtzev in ALA) has been reidentified to P. pedersenii (sect. Rubricaules), leaving the North American presence of P. lyngei subsp. lyngei in question. This is not the case for P. lyngei subsp. spissa Soják, described from Greenland; as discussed elsewhere (B. Ertter et al. 2013), this name is probably based on hybrids between P. pulchella and P. hyparctica (sect. Aureae), for which P. ×safronoviae Jurtzev & Soják is the priority name.

Additional nomenclatural details of species in sect. Pensylvanicae were provided by M. L. Fernald (1935), J. Soják (1987c), and B. Ertter (2008). For species of presumed hybrid origin with sect. Pensylvanicae as one parent, see sect. Rubricaules and Ertter et al. (2013).

(Discussion copyrighted by Flora of North America; reprinted with permission.)

Species ca. 400 (98, including 1 hybrid, in the flora).

Potentilla as presented here has a circumscription significantly reduced from treatments in most recent floras, since the convergence of morphologic (J. Soják 1985[1989], 2008) and molecular (T. Eriksson et al. 1998, 2003; C. Dobeš and J. Paule 2010; M. H. Töpel et al. 2011) interpretations has justified the segregation of P. fruticosa, P. glandulosa and relatives, P. palustris, and P. tridentata into Dasiphora, Drymocallis, Comarum, and Sibbaldiopsis, respectively (A. Kurtto and Eriksson 2003; Ertter 2007). The generic assignment of P. anserina is problematic (Soják 2010); the species is retained here within Potentilla. The senior author has chosen to maintain Duchesnea, Horkelia, Horkeliella, and Ivesia as distinct genera. As a result, the generic arrangement used here nearly recapitulates that used by P. A. Rydberg (1898, 1908d), representing the last comprehensive treatment of Potentilleae in North America.

Species circumscriptions and nomenclature are also significantly revised, reflecting not only the first continental assessment of the genus since Rydberg but also the need to synthesize American and Eurasian approaches to the genus. Whereas P. A. Rydberg’s (1908d) revision has served as the basis for subsequent floristic works in North America, subject to significant lumping of species and realignment of genera (for example, D. D. Keck 1938; J. Clausen et al. 1940), Eurasians generally have used as their starting point a monograph by T. Wolf (1908), a contemporary of Rydberg. Rydberg was handicapped by limited access to type material in Europe, resulting in the misapplication of several names.

Even with the reduced generic circumscription and improved global perspective, Potentilla remains challenging. Key contributing factors are facultative pseudogamous agamospermy and hybridization (summarized by J. Clausen et al. 1940; B. Eriksen 1996), which result in everything from ephemeral F1 hybrids to fully stabilized species of hybrid origin, sometimes across sectional boundaries. In the face of such biosystematic complexity, this treatment uses the operational species concept for arctic species outlined by R. Elven et al. (1999), with the exception that variety has been used as the primary infraspecific rank for temperate species.

Several species have been recently described or resurrected from synonymy for use in the current treatment, and anomalous collections examined during this preparation probably represent additional novelties. The isolated mountain ranges and complex geology of the North American Cordillera in particular contains a rich trove of unresolved diversity. Collectors should be alert for sympatric species in these situations, and the possible hybrid origin of aberrant individuals or clones.

Sections that have been established for North American Potentilla [for example, by A. Nelson (in J. M. Coulter and Nelson 1909), O. A. Stevens 1959, B. C. Johnston 1985] are mostly based on a subset of validly published but unranked groups by P. A. Rydberg (1898, 1908). Species numbers for primarily Eurasian sections have been provided by J. Soják (pers. comm.). The sectional arrangement used here is both pragmatic and tentative; it also represents an implicit phylogenetic hypothesis. The arrangement reflects molecular analyses (C. Dobeš and J. Paule 2010; M. H. Töpel et al. 2011), in which the first four sections correspond to well-supported clades that diverge basal to the ivesioid clade and the remaining sections. The majority of species beyond the first six sections are mingled in a largely unresolved crown clade, informally termed the Potentilla core group by Dobeš and Paule, and the Argentea clade by Töpel et al. Their results also suggest an Asian origin for Potentilla in the Oligocene, multiple colonizations in North America, and rapid radiation of the core group beginning in the Pliocene.

Detailed morphology and anatomy in Potentilla are provided by P. A. Rydberg (1898) and T. Wolf (1908). With some exceptions, floral morphology is of less taxonomic value than are leaf division and leaflet incision, particularly of basal leaves, which renders specimens without them often impossible to identify correctly. Habit and inflorescence architecture are also important features, such that specimens with fully developed inflorescences are typically more confidently identifiable than those in early anthesis.

Vestiture, especially of petioles and abaxial leaflet surfaces, is of critical diagnostic importance. Four primary hair groups are recognized by B. Eriksen and B. A. Jurtzev (1999): straight (or curved), crispate (short, twisted hairs, called crisped in the current treatment), floccose (corresponding to cottony in FNA terminology), and glandular (mostly short-stipitate). We further divide straight hairs into short hairs 0.2 mm, which sometimes form an even underlayer to other hair types, and long hairs for all other straight to curved hairs. Because these five groups can occur in any combination, all are addressed separately in descriptions. Hair density is categorized as sparse, common (or moderately), abundant, or dense if obscuring the surface; rigidity of long hairs is categorized as soft, weak, or stiff.

In keys and descriptions, stem length includes inflorescence, hairs are eglandular, and leaves are basal unless otherwise indicated. Rosetted plants have a basal rosette of leaves from a largely unbranched caudex; tufted plants have multiple stems from a short-branched caudex; and matted plants have dense, spreading, multi-branched caudices. Stolons and stoloniferous are used for plants that sprawl on the ground with apparently solitary flowers arising from nodes of flagelliform stems, even though these may be more homologous to prostrate cymose inflorescences than true stolons. Flagelliform stems form short shoots at the nodes but usually are not otherwise branched. In plants with stems arising centrally from ephemeral basal rosettes, new basal rosettes for the following year often appear in fall and might be confused with persistent rosettes. The diagnostic character of marcescent whole leaves (in sects. Niveae and Rubricaules) is in addition to sheathing stipular leaf bases and petioles, which are also commonly present on plants lacking marcescent whole leaves. Transitions between the extremes of strictly palmate and strongly pinnate leaves are roughly quantified as a fraction of the leaf axis (petiole plus rachis) bearing leaflets. Cauline leaves consist of foliar and bractlike structures on the stem proximal to the proximalmost branch. Leaflet dimensions and teeth number are of the larger leaflets (central in palmate leaves, terminal or distal in pinnate leaves). Leaflet tooth length is measured on the distal edge of the largest and medium-sized teeth. The central portion of the leaflet between opposite tooth sinuses is called the undivided medial blade. Hairs projecting beyond the tips of the leaflet teeth are referred to as apical tufts. Pedicel length is measured from the flower to the subtending bract, even if this appears to be a mid-pedicel bract. Because pedicels at the proximal nodes are often significantly longer than elsewhere, pedicel length is primarily for the nonproximal nodes, with the proximalmost extremes noted separately. Curved pedicels are those that commonly become sigmoidally recurved in fruit; this is characteristic of sections with predominately prostrate inflorescences. Measurements of pedicels, epicalyx bractlets, hypanthia, and sepals are taken at flowering and fruiting but not in bud. Hypanthium diameter is an external measurement of pressed flowers. Although not otherwise indicated, petals of most species commonly have a basal patch of more intense color, which often coincides with ultraviolet nectar guides (N. Naruhashi and H. Ikeda 1999). Style features are based primarily on dried mature styles that are tardily deciduous. Rugose achenes are often smooth until mature.

(Discussion copyrighted by Flora of North America; reprinted with permission.)