Annuals, biennials, or perennials, 20–300 cm, glabrous or tomentose. Stems erect, ascending, or spreading, simple or branched. Leaves basal and cauline; petiolate or sessile; proximal blade margins often ± deeply lobed, (spiny in C. benedicta), distal ± smaller, often entire, faces glabrous or ± tomentose, sometimes also villous, strigose, or puberulent, often glandular-punctate. Heads discoid, disciform, or radiant, borne singly or in corymbiform arrays. Involucres cylindric or ovoid to hemispheric. Phyllaries many in 6–many series, unequal, proximal part appressed, body margins entire, distal parts expanded into erect to spreading, usually ± dentate or fringed, linear to ovate appendages, spine-tipped or spineless. Receptacles flat, epaleate, bristly. Florets 10–many; outer usually sterile, corollas slender and inconspicuous to much expanded, ± bilateral; inner fertile, corollas white to blue, pink, purple, or yellow, bilateral or radial, often bent at junction of tubes and throats, lobes linear-oblong, acute; anther bases tailed, apical appendages oblong; style branches: fused portions with minutely hairy nodes, distinct portions minute. Cypselae ± barrel-shaped, ± compressed, smooth or ribbed, apices entire (denticulate in C. benedicta), glabrous or with fine, 1-celled hairs, attachment scars lateral (with or without elaiosomes); pappi 0 or ± persistent, of 1–3 series of smooth or minutely barbed, stiff bristles or narrow scales. x = 8, 9, 10, 11, 12, 13, 15.
Introduced; Eurasia, n Africa, widely introduced worldwide.
Species ca. 500 (20 in the flora).
Taxonomic limits of Centaurea have been controversial. The genus has great morphologic diversity, and studies have revealed much cytologic (e.g., N. Garcia-Jacas et al. 1996) and palynologic (e.g., G. Wagenitz 1955) variation as well. During the nineteenth and twentieth centuries, various taxonomists attempted, with limited success, to divide Centaurea into smaller genera or workable infrageneric taxa. The relations of several satellite genera have been controversial as well.
Recent molecular phylogenetic studies (A. Susanna et al. 1995; N. Garcia-Jacas et al. 2000, 2001) have begun to clarify relationships within Centaurea and between Centaurea and other genera. These studies make it clear that Centaurea as traditionally defined is polyphyletic, and that generic boundaries should be realigned if monophyletic taxa are to be recognized. Some taxa traditionally included within Centaurea (e.g., the two native North American species, Centaurea americana and C. rothrockii) fall outside the redefined generic boundaries and are here treated in Plectocephalus. Others usually placed into segregate genera (e.g., Cnicus benedictus) are firmly nested within Centaurea. Because the type species of Centaurea (C. centaurium Linnaeus, an African species) falls outside the main lineage of the genus, a proposal has been made to conserve Centaurea with a different type species (W. Greuter et al. 2001), thereby maintaining the nomenclatural stability of most of the numerous species that do fall within the principal Centaurea clade.
Although several Centaurea species are widely established as members of the North American flora, and some of these are widely distributed invasive weeds, some of the taxa listed by J. T. Kartesz and C. A. Meacham (1999) are apparently waifs and not permanent members of the flora. These taxa are discussed informally immediately below.
Centaurea aspera Linnaeus (rough star thistle) is known from nineteenth-century collections from ballast piles in New York; it does not appear to be established as a member of the North American flora. It can readily be distinguished from the similar C. diluta: the phyllary appendages are divided into palmately radiating clusters of short spines.
Centaurea babylonica Linnaeus has been reported from California as a waif (F. Hrusa et al. 2002); apparently it is not established as a permanent member of the flora. It is a tall (to 3 m), yellow-flowered Centaurea with numerous heads clustered in spiciform arrays. The phyllaries are leathery, and appendages are absent or reduced to spines 1 mm or shorter.
Centaurea cineraria Linnaeus (dusty miller), an Italian species known from casual garden escapes (California, Maryland, New York), probably is not permanently established as a member of the North American flora. It is a perennial with pinnately or bipinnately divided, densely gray-tomentose leaves, usually solitary, radiant heads somewhat larger than those of C. stoebe, and purple corollas.
Centaurea eriophora Linnaeus, reported from California and Colorado (J. T. Kartesz and C. A. Meacham 1999), would key below to C. sulphurea. It differs from C. sulphurea in having densely arachnoid-tomentose involucres. The California report is based on an early twentieth-century collection from Los Angeles County (A. Davidson 2334, UC). It is unlikely that this species is permanently established as a member of the California flora. The Colorado report is apparently erroneous, referenced to a report of Jacea pratensis Lamarck (= Centaurea jacea Miller), a wholly different plant, in W. A. Weber and R. C. Wittmann (1992).
A population of Centaurea trichocephala M. Bieberstein ex Willdenow (featherhead or hairy-head knapweed) was found in the late 1970s in a degraded pasture in eastern Washington (B. F. Roché and C. T. Roché 1991). A weed-control program was instituted, and the plants were successfully eradicated. Although it is apparently not established anywhere in North America, C. trichocephala is listed as a noxious weed in Oregon. These plants resemble C. phrygia in having elongate, pectinate-fringed phyllary appendages. In C. trichocephala the linear-filiform, featherlike appendages are much narrower than the phyllary bodies. Plants of the species spread by horizontal roots. According to Roché and Roché, C. trichocephala is apparently self-sterile; the Oregon plants spread clonally and formed no seeds.
Although Cnicus has usually been recognized as a distinctive monotypic genus, it has been merged into Centaurea by various authors (e.g., K. Bremer 1994; G. Wagenitz and F. H. Hellwig 1996). Recent molecular systematic studies (N. Garcia-Jacas et al. 2000) provide additional evidence that it is nested within Centaurea.
Centaurea bovina. J. T. Kartesz and C. A. Meacham (1999) listed C. bovina Velenovský from Massachusetts, referencing Rhodora 1924, perhaps based on collections reported as C. diffusa from Norfolk County (C. H. Knowlton and W. Deane 1924). Knowlton and Deane made no mention of C. bovina. According to J. Dostál (1976), C. bovina differs from C. diffusa in having smaller involucres (6–7 mm versus 7–10 mm; 3.5 mm versus 4–5 mm diameter), and purple versus pink flowers). Examination of one of the Norfolk County collections (Churchill s.n., MIN) revealed no differences from C. diffusa.
According to R. Angelo (pers. comm.), “The Kartesz citation is puzzling. This taxon [C. bovina] is not cited in the 1924 volume of Rhodora or anywhere in the first 50 years of Rhodora according to the index for that volume and our 50 year index (which is quite comprehensive).
“Also, there are no specimens identified as this taxon in the New England Botanical Club (NEBC) herbarium or from Massachusetts in the Harvard University Herbaria collections. We also looked for re-identifications in the other taxa in Centaurea in the NEBC and found no specimens that were earlier identified as C. bovina.
“The Vascular Plants of Massachusetts[...] (1999) of [B. A.] Sorrie and [P.] Somers does not list this taxon, nor does the Flora of the Northeast[...] (1999) by [D. W.] Magee and [H. E.] Ahles.”
Centaurea paniculata. According to our herbarium studies, reports of C. paniculata Linnaeus (Jersey knapweed) from North America are apparently referable to C. stoebe subsp. micranthos. Centaurea paniculata is quite similar to C. stoebe in habit; it differs clearly by its narrowly ovoid or cylindric heads.
Garcia-Jacas, N., A. Susanna, V. Mozaffarian, and R. Ilarslan. 2000. The natural delimitation of Centaurea (Asteraceae: Cardueae): ITS sequence analysis of the Centaurea jacea group. Pl. Syst. Evol. 223: 185–199.
Roché, B. F. and C. T. Roché. 1991. Identification, introduction, distribution, ecology, and economics of Centaurea species. In: L. F. James et al., eds. 1991. Noxious Range Weeds. Boulder, San Francisco, and Oxford. Pp. 274–291.
|1||Corollas yellow||> 2|
|1||Corollas white to pink, blue, or purple||> 6|
|2||Phyllary appendages broadly scarious, ± covering phyllary bodies, lacerate fringed, spineless or tipped by weak spines 1–2 mm||Centaurea macrocephala|
|2||Phyllary appendages spiny fringed, not covering phyllary bodies, tipped with spines 5–25 mm||> 3|
|3||Heads sessile, each closely subtended and ± concealed by involucrelike cluster of expanded, foliar bracts||Centaurea benedicta|
|3||Heads evidently pedunculate, or if sessile, each not concealed by involucrelike cluster of expanded, foliar bracts||> 4|
|4||Corollas usually longer than 25 mm||Centaurea sulphurea|
|4||Corollas usually 10–20 mm||> 5|
|5||Central spines of principal phyllaries 5–10 mm||Centaurea melitensis|
|5||Central spines of principal phyllaries 10–25 mm||Centaurea solstitialis|
|6||Principal phyllaries tipped with spines||> 7|
|6||Principal phyllaries not spine-tipped||> 11|
|7||Central spines of principal phyllaries 10–25 mm||> 8|
|7||Central spines of principal phyllaries usually 1–5 mm||> 9|
|8||Young leaves ± gray tomentose when young; involucre bodies 6–8 mm diam.; florets 25–40||Centaurea calcitrapa|
|8||Young leaves minutely bristly, green; involucre bodies 8–14 mm diam.; florets many||Centaurea iberica|
|9||Involucres 8–15 mm diam||Centaurea diluta|
|9||Involucres 3–5 mm diam||> 10|
|10||Corollas 12–13 mm; involucres 10–13 mm||Centaurea diffusa|
|10||Corollas 7–9 mm; involucres 7–8 mm||Centaurea virgata|
|12||Proximal leaves lanceolate, acute; phyllary appendages white to dark brown or black, teeth ± 1 mm; pappi 2–4 mm||Centaurea cyanus|
|12||Proximal leaves oblong, obtuse; phyllary appendages silvery white to brown, teeth 1.5–2 mm; pappi 6–8 mm||Centaurea depressa|
|13||Phyllary appendages tapering to long, often recurved, pectinately dissected, filiform tips||Centaurea phrygia|
|13||Phyllary appendages obtuse to acute, erect or ascending||> 14|
|14||Involucres 10–13 mm||Centaurea stoebe|
|14||Involucres 15–25 mm||> 15|
|15||Phyllary appendages evidently decurrent along phyllary margins||> 16|
|15||Phyllary appendages not or only slightly decurrent along phyllary margins||> 17|
|16||Leaves decurrent, entire or remotely dentate, or proximal some- times pinnately lobed||Centaurea montana|
|16||Leaves not decurrent, 1–2-pinnately divided or distal entire.||Centaurea scabiosa|
|17||Phyllary appendages roundish, seldom triangular, scarious, light to dark brown, ± undivided to irregularly lacerate||> 18|
|17||Phyllary appendages ± triangular, brown to black, ± wholly pectinately fimbriate; pappus present or absent||> 19|
|18||Pappus absent||Centaurea jacea|
|18||Pappus present, 3–5 mm, of many white bristles||Centaurea diluta|
|19||Heads discoid; peripheral florets not expanded, showy; pappus blackish, shorter than 1 mm; green parts of phyllaries ± totally covered by black appendages, involucres appearing totally black||Centaurea nigra|
|19||Heads radiant; peripheral florets expanded, showy, raylike; pappus absent or rudimentary, when present usually not black; green part of phyllaries sometimes evident or appendages light to dark brown||> 20|
|20||Heads relatively broad, the pressed involucres usually as wide as or wider than long; green parts of phyllaries usually fully covered by ± brown, variably pectinately fimbriate appendages, involucres light to dark brown||Centaurea ×moncktonii|
|20||Heads relatively narrow, the pressed involucres usually longer than wide; green parts of phyllaries not fully covered by black appendages, involucres black and green||Centaurea nigrescens|