Taxonomic revision of Echinacea

We sampled wild populations of each of McGregor's (1968) Echinacea taxa and performed a numerical and cladistic analysis of variation using morphological (and some chemical) characteristics (Binns et al., 2002a). Natural populations were tentatively identified in the field according to McGregor's taxonomy (1968). Voucher specimens were deposited at the Department of Agriculture Ontario Herbarium (Ottawa, Canada) and experimental plants and seed were grown in a greenhouse (Binns et al., 2002a, 2002b, 2002c). Measurements for 81 morphometric traits (binary, quantitative, semiquantitative, and qualitative) were entered in a matrix to determine the degree of relationship and clustering between specimens, without any a priori weighting according to previous taxonomic identification. The data for seven traits were omitted due to missing values, resulting in a data matrix of 321 individuals by 74 characters (traits). An index of overall similarity was calculated for each pair of individuals (Gower, 1971). Several clustering methods were used to determine close relationships and potential taxonomic clusters (see Binns et al., 2002a, for details). Clusters generated in these analyses represented potential taxonomic entities. We evaluated clusters using canonical discriminant analysis (Kshirsagar, 1972) to determine the relative strength of different cluster solutions. We also took the limitations of each clustering method into consideration, such as bias of equal-sized clusters, equal variance for members of a cluster, and misrepresentation of relationship between clusters due to error associated with extreme linkage methods. Herbarium specimens were also analyzed without a priori identification, including many from McGregor's collection. Finally, we also tested McGregor's taxonomy by using his keys to identify and label each specimen in our matrix of similarities, and then applying clustering strategies and assessing the validity of these solutions.

The morphometric analyses supported two acceptable cluster solutions based on overall similarity between specimens. The first revealed strong statistical support for two major divergent taxa within Echinacea, which we determined to be at a subgenus level. The species known currently as E. purpurea (L.) Moench was the sole taxon in Echinacea subgenus Echinacea, which contains only the species E. purpurea, and all other infrageneric taxa were in Echinacea subgenus Pallida. The second most acceptable cluster solution supported four taxa, which we determined to be at the species level: E. purpurea [=Echinacea purpurea (L.) Moench nom. cons. prop.; see Binns et al., 2001b], E. laevigata [=E. laevigata (Boynton & Beadle) Blake], E. atrorubens and E. pallida. There was also significant statistical support for an eight-cluster solution based on a priori labeling of each specimen using McGregor's (1968) taxonomy. The eight groups correspond to varieties within two species. The varieties of E. pallida follow: (1) E. pallida var angustifolia; (2) E. pallida (Nutt.) Nutt. var. simulata (McGregor) Binns, B.R. Baum & Arnason [=E. simulata McGregor]; (3) E. pallida (Nutt.) Nutt. var. tennesseensis (Beadle) Binns, B.R. Baum & Arnason [=E. tennes-seensis (Beadle) Small]; (4) E. pallida (Nutt.) Nutt. var. sanguinea (Nutt.) Gandhi & Thomas [=E. sanguinea Nutt.]; and (5) the type variety, E. pallida var. pallida. The varieties of E. atrorubens include (1) E. atrorubens Nutt. var. paradoxa (Norton) Cronq. [=E. paradoxa (Norton) Britton var. paradoxa]; (2) E. atrorubens Nutt. var. neglecta (McGregor) Binns, B.R. Baum & Arnason [=E. paradoxa (Norton) Britton var. neglecta McGregor]; and (3) the type variety, E. atrorubens Nutt. var. atrorubens Cronq. [=E. atrorubens Nutt].

Echinacea subg. Echinacea has the following key characteristics that distinguish between the two subgenera: fibrous roots, basal leaves usually 50 to 100 mm wide (rarely to 150 mm) and cauline leaves usually 45 to 90 mm wide, leaf blade trichomes bicellular with basal cell cylindrical and distal cell acuminate, branched major leaf veins, and an involucre with four series of bracts. Echinacea subg. Echinacea is distributed in sparse natural populations throughout the eastern and southeastern United States in a range that overlaps geographically, but rarely ecologically, with all species in Echinacea subg. Pallida. In Figure 1.1, a map of the distribution of each revised species, subg. Echinacea is shown as the range of E. purpurea (Binns et al., 2002a). For descriptions of each revised species and variety, as well as two morphological identification keys and other distribution maps, see Binns et al. (2002a).

The evolutionary relationships between the four species were estimated using cladistic analyses of 36 characters (including some phytochemical ones). In Figure 1.2, the monophyletic clade Echinacea is distinct from the outgroup Rudbeckia (98% bootstrap value). Within the Echinacea clade, the relationship was closest between E. atrorubens and E. pallida, which share three unique, derived characteristics, and E. purpurea was most basally divergent. The taxon E. laevigata was more closely related to E. pallida and E. atrorubens, based on the current data, although historically it was confused with E. purpurea.

In summary, we proposed a hierarchy of two subgenera, four species, and six varieties in the genus Echinacea from our morphometric numerical analyses (Binns et al., 2002a). The two subgenera are novel, but the four species groups were first suggested based on intuitive taxonomic methodology (Cronquist, 1945), and taxa at the variety level were all previously either species or varieties according to McGregor (1968). The revised taxonomy recognizes all of McGregor's taxa, except for one variety, E. angustifolia DC. var. strigosa McGregor. This putative variety may be a morphotype established from introgressed hybrids with similar phenotypic development in similar ecological zones, but not found to be distinct from E. pallida var. angustifolia (Binns, 2001). The

FIGURE 1.1 Distribution of four revised Echinacea species (including their varieties) from a combination of modern collections and herbarium specimens. Contour lines indicate predicted occurrence based on historical data as well as current surveys at the time of publication. (From Binns, S.E. et al., Syst. Bot., 27(3): 610-632, 2002a. With permission.)

FIGURE 1.1 Distribution of four revised Echinacea species (including their varieties) from a combination of modern collections and herbarium specimens. Contour lines indicate predicted occurrence based on historical data as well as current surveys at the time of publication. (From Binns, S.E. et al., Syst. Bot., 27(3): 610-632, 2002a. With permission.)

|—n—Rudbeckia sect. macrocline ■— R. sect. laciniata

R. sect. rudbeckia

■— Echinacea purpurea -Q— E. laevigata (70%) — E. pallida

FIGURE 1.2 A 40-step most parsimonious cladogram representing the monophyletic genus Echinacea Moench compared to three sections of Rudbeckia in an outgroup. Bootstrap values using the 50% majority-rule consensus method are indicated in brackets below the branches. Cladistic analysis was performed with 36 characters. Dark boxes signify synapomorphies and empty boxes signify parallelisms. (Modified version from Binns, S.E. et al. Syst. Bot., 27(3): 610-632, 2002a. With permission.)

classifications of both McGregor (1968) and Cronquist (1945, 1955, 1980) are compared to the revised taxonomy in Binns et al. (2002a).

Was this article helpful?

0 0
Building Your Own Greenhouse

Building Your Own Greenhouse

You Might Just End Up Spending More Time In Planning Your Greenhouse Than Your Home Don’t Blame Us If Your Wife Gets Mad. Don't Be A Conventional Greenhouse Dreamer! Come Out Of The Mould, Build Your Own And Let Your Greenhouse Give A Better Yield Than Any Other In Town! Discover How You Can Start Your Own Greenhouse With Healthier Plants… Anytime Of The Year!

Get My Free Ebook

Post a comment