Flora of North America species comparison

Descurainia nelsonii was treated by L. E. Detling (1939) and R. C. Rollins (1993) as a subspecies of D. pinnata, but the latter in the sense of these authors is not monophyletic, comprising instead either four or two unrelated species, respectively. ITS molecular data (B. E. Goodson 2007) suggest that D. nelsonii is most closely related to D. longepedicellata and D. paradisa. It can be distinguished from the latter species by its linear fruits with cuneate tips; D. paradisa has obovoid fruits with rounded tips. Descurainia nelsonii resembles D. pinnata subsp. brachycarpa in the orientation of fruiting pedicels and in having short styles (to 0.3 mm) and small seeds (to 1 × 0.5 mm). It differs in being branched (versus simple) at base and in having smaller flowers (petals 0.7–1 versus 1.5–2.6 mm), fewer ovules (6–12 versus 16–40) per ovary, linear (versus subclavate) fruits, and uniseriate (versus biseriate) seeds.

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

Subspecies 4 (4 in the flora).

Plants assigned to Descurainia pinnata by various authors represent one of the most complex assemblages of any North American Brassicaceae. L. E. Detling (1939) divided it into eleven subspecies, of which R. C. Rollins (1993) accepted eight. The species includes series of populations that show tremendous variability in every conceivable character, and the variation is continuous between extremes. Some of the taxa recognized by these authors were based solely on trivial variations in the degree of pubescence on the distal parts, shape of the distalmost leaf segments, and presence or absence of glands. The latter character can vary within a given population and, therefore, its utility in distinguishing taxa is questionable at best. Cases in point are Cully s.n. (spring 1984) and Worthington 26353, both at UNM, collected from Bernalillo and Luna counties (New Mexico), respectively; each sheet has both glandular and eglandular forms.

As delimited by L. E. Detling (1939) and R. C. Rollins (1993), subspp. halictorum and menziesii represent a heterogeneous assemblage of intermediates within Descurainia pinnata and between it and other species. In the absence of thorough studies on these subspecies, we prefer to avoid placing them in the synonymy of a putative parental species or recognizing them as subspecies of a parent. Readers interested in the synonymies of subspp. halictorum and menziesii may consult Detling and Rollins.

Molecular studies by B. E. Goodson (2007) indicated that some taxa previously included in Descurainia pinnata (i.e., subspp. filipes, nelsonii, paradisa, and paysonii) should be placed elsewhere. While the molecular data were not able to resolve the remaining subspecies, a critical evaluation of morphology shows that the entire D. pinnata complex falls into at least four relatively distinct groups recognized here as subspecies.

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