Flora of North America species comparison

Of conservation concern.

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

Genera ca. 1500, species ca. 23,000 (418 genera, 2413 species in the flora).

Asteraceae (Compositae, “composites,” or “comps”) have long been recognized as a natural group, and circumscription of the group has never been controversial (although some authors have divided the traditional family into three or more families). A. Cronquist (1981) placed Asteraceae as the only family in the order Asterales within subclass Asteridae, associated with the Gentianales, Rubiales, Dipsacales, and Calycerales and relatively distant from Campanulales. On recent molecular phylogenetic data, the Angiosperm Phylogeny Group (2003; see references there for details; classification abbreviated APGII hereafter) has suggested that Asteraceae are better treated as part of a more widely defined Asterales within the asterids II informal clade (or campanulid clade; see W. S. Judd and R. G. Olmstead 2004). Judd and Olmstead summarized the higher-order relationships of Asteraceae as follows (in order of decreasing inclusiveness; synapomorphies in parentheses): asterids (ovules unitegmic and tenuinucellate, iridoid chemistry); core asterids (sympetaly, stamen number equal to petal number, stamen epipetaly, mostly 2–3-carpellate gynoecia); campanulids (early sympetaly), comprising eight unassigned families plus Aquifoliales, which is sister to Dipsacales, Apiales, and Asterales (last three sharing frequently inferior ovaries, polyacetylenes); and Asterales, which appears to be sister to Dipsacales-Apiales (K. Bremer et al. 2004). The order Asterales (valvate petals, lack of apotracheal parenchyma, storage of inulin, ellagic acid present, and, possibly, the presence of a plunger or brush pollen presentation mechanism) now includes the following families (fide APGII): Alseuosmiaceae, Argophyllaceae, Calyceraceae, Campanulaceae (optionally including Lobeliaceae), Goodeniaceae, Menyanthaceae, Pentaphragmaceae, Phellinaceae, Rousseauaceae, and Stylidiaceae. Within Asterales, Asteraceae is part of a clade (corollas with more or less fused lateral veins joining midvein near lobe apices, thick integuments, no endosperm haustorium) with the Menyanthaceae (cosmopolitan with Southern Hemisphere genera) basal to a more nested clade (inferior ovaries, possibly connate anthers, pollen exine with bifurcating columellae) comprising Asteraceae, Goodeniaceae (mainly Australia), and Calyceraceae (South America), the last being the immediate sister to Asteraceae (highly modified, persistent calyces, corolla venation patterns, unilocular and uniovulate gynoecia, pollen with intercolpar depressions, specialized fruits). Aggregation of flowers into heads with involucres appears to have been a parallel phenomenon in Calyceraceae and Asteraceae, given the determinate nature of the former and indeterminate (racemose) organization of the latter. Some traits typical of Asteraceae predate evolution of the family as a distinct clade. Relationships of Asteraceae and Calyceraceae have been discussed by M. H. G. Gustafsson and Bremer (1995). Synapomorphies of the Asteraceae clade include: calyces modified to structures called pappi, anthers connate (forming tubes) and styles modified to function as brushes in a specialized pollen presentation mechanism, ovaries each containing a single basal ovule, and production of sesquiterpene lactones.

K. Bremer et al. (2004) gave an Early Cretaceous origin for the Asteridae and the basal campanulids, and a Late Cretaceous origin for the Asterales. Bremer and M. H. G. Gustafsson (1997) also hypothesized a Late Cretaceous ancestry of Asterales in East Gondwanaland (Australasia), with later expansion into West Gondwanaland (South America-Antarctica), where the Asteraceae originated before the final separation of South America and Antarctica. Similarly, M. L. DeVore and T. F. Stuessy (1995) argued that the close relationships of Asteraceae to Goodeniaceae and Calyceraceae, plus the basal position of Barnadesioideae K. Bremer & R. K. Jansen (Asteraceae), indicated a South America-Antarctica-Australia origin for the complex. After reviewing previous hypotheses, they proposed a late Eocene origin for the complex and suggested a South American origin for the Asteraceae based on the basal position of the South American Barnadesioideae (see also Stuessy et al. 1996, on Barnadesioideae origin in southern South America in the Oligocene) and their sister relationship to Calyceraceae. Fossil pollen data (both Mutisieae and Asteroideae types—notably Heliantheae in the broad sense—among earliest reports) reviewed by A. Graham (1996) appear to indicate an Eocene origin for Asteraceae in South America, with migration to North America at least by the Oligocene, possibly as early as the late Eocene. More recently, M. S. Zavada and S. E. de Villiers (2000; and references therein) reported Asteraceae pollen (assignable to Mutisieae in the broad sense) from the Paleocene-Eocene of South Africa, suggesting an earlier, West Gondwana (southern Africa or Australia) origin for the family. Such data indicate that some tribes of Asteraceae may have arrived in North America via long-distance dispersal or island hopping well before closure of the isthmus of Panama. They also have a bearing on the possible times of radiation of some tribes in North America, particularly Heliantheae in the broad sense and Eupatorieae, which originated in the continent (including Mexico and parts of Central America), and those that came to North America from or through South America such as Mutisieae, Vernonieae, some Plucheeae, and Astereae. Other tribes, such as Cynareae, Cichorieae, some Gnaphalieae, and Anthemideae, may have reached North America from Eurasia, possibly via Beringia (or as Amphi-Atlantic disjuncts), at a later time.

The bases of a tribal classification within Asteraceae were established in the nineteenth century, primarily through the work of H. Cassini (especially in articles scattered through the 61 volumes of F. Cuvier 1816–1845; Cassini included synopses of his tribes as part of his entry for Zoegea, i.e., zyégée in French; the articles have been collected in three volumes by R. M. King and H. W. Dawson 1975), C. F. Lessing (1832), A. P. de Candolle (1828–1838, 1836–1838), and, particularly, G. Bentham (1873). In the twentieth century, the tribal system of Cassini, as elaborated by Bentham, was widely followed with only slight modifications (see S. Carlquist 1976; A. Cronquist 1955, 1977; C. Jeffrey 1978; G. Wagenitz 1976b; see also J. Small 1919 and, for alternate views on Heliantheae-Eupatorieae, H. Robinson 1996).

A molecular phylogenetic study by R. K. Jansen and J. D. Palmer (1987) established that a South American clade (later named Barnadesioideae) is basal within Asteraceae. Both cladistic morphologic analyses (e.g., K. Bremer 1994, 1996) and mostly chloroplast-DNA molecular phylogenies (e.g., Jansen et al. 1991, 1992; K. J. Kim et al. 1992; Kim and Jansen 1995; R. J. Bayer and J. R. Starr 1998; P. K. Eldenäs et al. 1999; B. G. Baldwin et al. 2002) have deepened our knowledge of tribal interrelationships within Asteraceae and led to the recent proposal of a phylogenetic classification for the family with 10 subfamilies and 35 tribes (J. L. Panero and V. A. Funk 2002).

Treatment of Asteraceae here differs from some of the recently proposed classifications in that some groups continue to be traditionally circumscribed (e.g., Mutisieae in the broad sense, Heliantheae in the broad sense, including Helenieae and excluding Eupatorieae). Where appropriate and so far as practicable, new taxonomies are acknowledged in our discussions of individual tribes (which see). In North America, the following subfamilies and tribes, as defined by J. L. Panero and V. A. Funk (2002), are represented (tribes with no native representatives are marked by asterisks): Mutisioideae-Mutisieae in the strict sense, Gochnatioideae-Gochnatieae, and Hecastocleioideae-Hecastocleideae (all included in Mutisieae here, which see), Carduoideae (Cardueae = Cynareae), Cichorioideae (*Arctoteae, Cichorieae, Vernonieae), and Asteroideae [Senecioneae, *Calenduleae, Gnaphalieae, Anthemideae, Astereae, Plucheeae, *Inuleae, Eupatorieae, and the following segregates of Heliantheae in the broad sense (all treated here within or as subtribes of a fairly traditionally circumscribed Heliantheae): Bahieae, Chaenactideae, Coreopsideae, Helenieae, Heliantheae in the strict sense, Madieae, *Millereae, Perityleae, Polymnieae, and Tageteae)].

Asa Gray produced the first broadly influential floristic synthesis of North American Asteraceae. Other authors who made important contributions to floristics of North American Asteraceae in the nineteenth and first half of the twentieth centuries were S. F. Blake, N. L. Britton, R. S. Ferris, M. L. Fernald, E. L. Greene, H. M. Hall, M. E. Jones, D. D. Keck, P. A. Rydberg, J. K. Small, and S. Watson. Some of those authors had narrower concepts of genera and species than had their predecessors and they freely recognized new taxa in Asteraceae (mostly genera and species). Floristics of North American Asteraceae in the second half of the twentieth century was especially influenced by A. Cronquist (e.g., 1955, 1980, 1994; H. A. Gleason and Cronquist 1991), who usually favored traditional generic circumscriptions.

In the last 20 years or so, developments in molecular systematics have led to revisions of generic limits in some tribes of Asteraceae and, sometimes, to a return to generic concepts that had been suggested earlier but largely ignored. More or less worldwide, taxonomies in some tribes or parts of tribes have included segregate genera that have been revived or newly published. Most of the innovations will be summarized in the forthcoming Asterales volume of K. Kubitzki et al. (1990+). The generic circumscriptions adopted here incorporate recent taxonomic findings relevant to North America, insofar as our contributors have accepted them. As a result, many of the genera treated herein have never been presented in a major flora before, and some species are included within genera with which they were not associated traditionally. Thus, the Flora brings together much new knowledge and many new names. In most instances, circumscriptions of species have turned out to be conventional. So far as practicable, recently named species from North America have been accounted for within relevant treatments herein.

With 418 genera and 2413 species (Table 1), Asteraceae is, numerically, the largest family in the flora of North America north of Mexico. Members of the family are found in diverse habitats, from the High Arctic tundra and polar deserts to the Sonoran warm-desert scrub, and from alpine habitats to salt marshes. Asteraceae are particularly conspicuous elements of warm-desert and intermountain grasslands, as well as of desert scrubs, notably the intermountain desert scrub where Artemisia dominates (M. G. Barbour and N. L. Christensen 1993). Among other conspicuous species, members of Solidago and Symphyotrichum form a very showy part of the fall flowering in eastern North America, and members of Heliantheae sometimes produce striking displays in the American West (e.g., Gaillardia spp., Lasthenia spp., members of Madiinae).

Much has been published, not only on systematics (at various levels), but on biology, chemistry, and economic and medical uses of Asteraceae worldwide, particularly in proceedings (from conferences and symposia) edited by V. H. Heywood et al. (1977), T. J. Mabry and G. Wagenitz (1990), and D. J. N. Hind et al. (1995, 1996).

Relatively few North American species of Asteraceae are economically important or widely used ethnobotanically. The only major Asteraceae crop of North American origin is the sunflower, Helianthus annuus, which is valued for its seed oil and is appreciated in the horticultural trade. Other crop plants from native species worth mention are Helianthus tuberosus, the Jerusalem artichoke, and Parthenium argentatum, the guayule, a source of rubber. Echinacea spp. are touted as health plants. Members of several genera of Asteraceae native to the flora are grown for their ornamental value, notably species of Coreopsis (tickseeds), Echinacea (coneflowers), Helianthus (sunflowers), Liatris (blazingstars and gayfeathers), Rudbeckia (black-eyed Susans), Solidago (goldenrods), and Symphyotrichum (“asters” of the trade).

Many species of Asteraceae have been introduced into North America, mainly from Europe and Asia, some deliberately for medicines, foods, or horticulture, others accidentally (often with seeds or other agricultural products or by other means). Few, if any, of the introduced taxa are thought to be noxious at the continental level, but some (e.g., Acroptilon) are considered noxious in large parts of their ranges within the flora. Taraxacum officinale is a common lawn weed that (in terms of dollars spent and herbicides applied in weed control) has an economic and ecologic impact disproportionate to the actual harm it causes; other weedy introduced Asteraceae are of little economic consequence. Some native Asteraceae are toxic to cattle and other livestock and are therefore considered weeds. And some native species of open habitats (e.g., Symphyotrichum pilosum) are often considered weeds because they invade fields left fallow. Ragweeds (especially Ambrosia artemisiifolia and A. trifida) range over nearly the whole continent and their wind-blown pollens cause late–summer allergic reactions (hayfever) for a large number of people. Because ragweeds have a large impact on human health, they have a significant, negative economic impact.

In contrast to Orchidaceae, for which a wealth of excellent, well-illustrated popular books are available, few popular field guides on Asteraceae of North America have been published. The guide by T. M. Antonio and S. Masi (2001) deserves notice for its maps, color photographs, and useful information.

Composites (members of Asteraceae) share some unusual morphologic traits and some morphologic terms are used in particular ways as applied here to them.

For treatments of composites here, “perennials” are herbaceous and differ from annuals and biennials in living longer than two years and differ from subshrubs, shrubs, and trees in not developing woody aerial stems.

In most composites, leaf venation comprises a midrib plus more or less equal lateral nerves or veins; such leaves are described as pinnately nerved. Venation in leaf blades of some composites often consists of a midrib plus relatively strong lateral veins that diverge at or just distal to bases of blades. Such leaves are described as 3-nerved, 3(–5)-nerved, 5-nerved, etc., and, as appropriate, the phrases “from bases” or “distal to bases” may be added for clarification.

Composites often have subsessile to sessile or sunken glandular hairs that consist of multicellular bases supporting globular elements that usually contain resinous or sticky substances. Such structures have been called glands, glandular hairs, glandular trichomes, punctae, resin dots, and so on. Sometimes, the glands are embedded in epidermal depressions or pits. Epidermes with glands more or less sunk into or embedded within the surface have been called glandular-punctate and/or punctate-glandular. The glands may be colorless (translucent) or yellowish to dark brown or orange and are sometimes more prominent on dried specimens than in living plants. In keys and descriptions here, gland-dotted refers to the presence of such glandular hairs, whether sessile or in depressions or pits (as appropriate, “in pits” or “sessile” may be added for clarification).

Inflorescences of composites are called heads (or capitula, sing. capitulum). Heads may be borne singly (i.e., not clearly associated with other heads on the same plant) or associated in arrays. The arrays of heads on composites correspond to arrays of individual flowers (inflorescences) on plants of other families; arrays of heads are sometimes called capitulescences. Terms for architectural structures of arrays of heads are parallel to terms for kinds of inflorescences: cymiform, corymbiform, paniculiform, racemiform, spiciform, thyrsiform, etc.

In radiate heads, peripheral florets (ray florets) in one or more series have corollas with zygomorphic limbs and may be pistillate, or styliferous and sterile, or neuter; the central florets (disc florets) in radiate heads have ± actinomorphic corollas and may be bisexual or functionally staminate. In liguliflorous heads, all florets are bisexual and (usually) fertile and have zygomorphic corollas (ligulate florets); liguliflorous heads are characteristic of Cichorieae and are found in no other composites. In discoid heads, all florets have ± actinomorphic corollas and all are either bisexual and fertile or all are either functionally staminate or pistillate (in monoecious or dioecious taxa, e.g., Baccharis spp.). In disciform heads, all florets have ± actinomorphic corollas, and peripheral florets (in one or more series) are usually pistillate and usually have relatively slender (often filiform) corollas. Such peripheral pistillate florets are generally thought to be derived by reduction from ray florets, and plants with disciform heads are generally thought to be derived from ancestors with radiate heads. The central florets of disciform heads are usually bisexual, sometimes functionally staminate. By tradition and for simplicity, both the peripheral, pistillate florets and the inner, bisexual or functionally staminate florets in disciform heads may be referred to as “disc” florets. In radiant heads, all florets have ± actinomorphic corollas and the peripheral florets usually have much enlarged corollas and may be bisexual, pistillate, or neuter; the central florets of radiant heads are usually bisexual. Some composites have peripheral, bisexual florets with slightly to strongly zygomorphic corollas (e.g., some members of Chaenactis, Lessingia, Thymophylla, et al.); heads of such plants do not quite conform to any of the five types just described and such heads may be referred to as “quasi-radiate” or “quasi-radiant.” Some florets in heads of some Mutisieae have 2-lipped corollas and those heads may be called “quasi-radiate” or “quasi-liguliflorous.” The term eradiate is used to refer collectively to discoid, disciform, and radiant heads.

Heads with all florets of one sexual form (bisexual, pistillate, or functionally staminate) are called homogamous (discoid and liguliflorous heads are homogamous, some radiant heads may be homogamous) and heads with florets of two or more sexual forms are called heterogamous (radiate and disciform heads are heterogamous, some radiant heads may be heterogamous).

Phyllaries collectively constitute an involucre, usually number 5–21(–50+), usually are unequal (outermost usually shorter than the inner), and usually are arranged ± imbricately (overlapping like shingles) in 3–5(–15+), usually ± spiral series. Sometimes, the phyllaries are ± equal in 1–2 series; they are rarely wanting (e.g., Psilocarphus spp.). Phyllaries may be herbaceous or chartaceous to scarious and are often medially herbaceous with chartaceous to scarious borders and/or apices. The phyllaries “proper” are sometimes immediately subtended by a calyculus (pl. calyculi) of (1–)3–15+ distinct, usually shorter bractlets in 1(–3+) series (e.g., Coreopsis spp., Taraxacum spp.).

Receptacles may bear paleae (i.e., some or all florets are individually subtended by a bractlet called a palea or receptacular bract). Collectively paleae have been called “chaff” and paleate receptacles have been described as “chaffy.” Receptacles that bear paleae are referred to as paleate and receptacles that never bear paleae are referred to as epaleate. Epaleate receptacles sometimes bear subulate enations (e.g., some Gaillardia spp.) or bristles or subulate to linear scales (e.g., some Cynareae), or fine hairs (e.g., some Anthemideae). Epaleate receptacles (and paleate receptacles that have shed their paleae) may be smooth or pitted (alveolate, foveolate, etc.).

The terms tube, throat, and limb have been variously used in descriptions of corollas of composites. Here, in ± actinomorphic corollas of bisexual and functionally staminate disc florets, the tube is the part of the corolla proximal to the insertion of the staminal filaments, and the limb is the part that is distal to insertion of the filaments. The limb comprises, proximally, the throat and, distally, the lobes. The distinction between tube and throat hinges on insertion of filaments, not on external morphology.

The relatively flat portion of a corolla of a ligulate floret from a liguliflorous head (i.e., members of Cichorieae) is called a ligule; it terminates in 5 teeth or lobes. The relatively flat portion of a corolla of a ray floret is called a lamina; it terminates in 0–3(–4) teeth or lobes. More or less bilabiate corollas are characteristic of some members of Mutisieae and are seldom found in members of other tribes.

Fruits of composites have been called “achenes” because they resemble true achenes. Achenes are dry, hard, single-seeded fruits derived from unicarpellate, superior ovaries. Ovaries of composites are bicarpellate and inferior. Fruits derived from ovaries of composites are called cypselae (sing. cypsela, a term coined by C. de Mirbel in 1815). Morphology of an ovary of a composite at flowering is often markedly different from the morphology of the mature fruit (cypsela) derived from that ovary. References to cypselae in keys and descriptions here almost always refer to mature fruits, not to ovaries at flowering.

Shapes of cypselae have been used in distinguishing among species, genera, and even subtribes of composites. In most composites, cypselae are ± isodiametric in cross section. In some composites, cypselae are characteristically ± lenticular to elliptic in cross section. Such cypselae are said to be compressed (or laterally flattened) if the longer axis of the cross section is ± parallel to a radius of the head (e.g., Verbesina spp.). Cypselae are said to be obcompressed (or radially flattened) if the shorter axis of the cross section is ± parallel to a radius of the head (e.g., Coreopsis spp.).

In composites, pappi (sing. pappus) are found where calyces are usually found on inferior ovaries; pappi have been shown to be greatly modified calyces. They show a great range of diversity and are often diagnostic for recognition of taxa, especially at rank of genus and below. The forms of individual pappus elements intergrade. For keys and descriptions here, the following distinctions are made: cross sections of bristles and awns are ± circular or polygonal and have the longer diameter of the cross section no more than 3 times the shorter diameter. Pappus elements with “flatter” cross sections (i.e., longer diameter more than 3 times the shorter diameter) are called scales, regardless of relative overall lengths and widths of the elements. As used here, “subulate scale” and “setiform scale” mean much the same as “flattened bristle” of some authors. Pliable to stiff pappus bristles with diameters less than ca. 50 mm are called fine bristles; pliable to stiff bristles with diameters 50–100 mm are called coarse bristles. Rigid pappus elements with ± circular or polygonal cross sections greater than 100 mm in diameter are called awns. Bristles, awns, and scales may be smooth or finely to coarsely barbed or plumose. A scale of a pappus may terminate in one or more bristlelike or awnlike appendages; such scales are said to be aristate.

In keys and descriptions, “pappus” and “pappi” usually refer to structures found on cypselae (mature fruits), not to “immature pappi” of ovaries at flowering. Sometimes pappi of ovaries that do not form fruits (e.g., in functionally staminate florets of some tarweeds) may be taxonomically useful and may be referred to in descriptions and keys.

Following is a synoptic key to tribes into which genera of composites of the flora area are placed. Keys to genera within each tribe will be found in the accounts of the individual tribes. Because some traits in the key to tribes and in keys to genera within tribes may be difficult to assess, we have also provided a key to artificial groups of composites and keys to genera within those artificial groups. Those keys will be found following the key to tribes.

In the following key, “radiate heads” have ray florets; “eradiate heads” lack ray florets and may be disciform, discoid, or radiant. Ray florets have zygomorphic corollas with laminae; the laminae may be showy (as in some species of Helianthus) or inconspicuous (as in some species of Erigeron). Usually, we have included plants with inconspicuous ray laminae in keys to genera of both radiate and eradiate groups.

Some plants have questionably paleate or epaleate receptacles. Epaleate receptacles of some plants are notably pitted and have fimbriate to deeply lacerate pit borders; such receptacles have sometimes been interpreted as paleate. Plants with notably lacerate pit borders are usually keyed here as both paleate and epaleate.

Some plants with pappi of conspicuous bristles often have the bristles subtended by minute, inconspicuous scales. Although such plants technically belong to groups with pappi “wholly, or partially, of awns or scales,” they are usually also keyed here in groups characterized as having pappi “wholly of bristles,” because the scales are easily overlooked. As well, some pappus elements are borderline between being called subulate or setiform scales or being called “flattened bristles.” Consequently, some plants that technically belong to groups with pappi of scales are keyed both in groups with pappi “wholly of bristles” and in groups with pappi “wholly, or partially, of awns or scales.”

Table 1: [see original page on floranorthamerica.org]

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