Flora of North America species comparison

Species ca. 230 (169, including 17 hybrids, in the flora).

Species of Crataegus vary from heliophiles to moderately shade tolerant. Most hawthorns are mesophytes and soil moisture regime appears to be a more important habitat determinant than soil type. Hawthorns are mostly found on topland, mesic, fine-textured soils, usually reflective of a reasonable rainfall. None are strict calciphiles or acidophiles. In drier parts of the Great Plains and intermontane regions, suitable habitats are often restricted to the vicinity of watercourses, or marsh edges, seeps, ditches, etc. Relatively few taxa grow in a drier (for example, ser. Lacrimatae) or moister (ser. Aestivales) soils than the norm.

Often abundant, hawthorns can be important components of some ecosystems, acting as sources of shelter and protection for animal life, nesting habitat, source of food (pomes), mainly for medium-sized passerine birds, and source of nectar and pollen. Hawthorns may be nurse plants to deciduous trees and play an important role in succession. Gymnosporangium R. Hedwig rusts may be serious pests on hawthorns where junipers are common. Other, locally serious, threats to native hawthorns include aggressive introduced woody competitors such as Lonicera japonica Thunberg in the southeast and Rhamnus cathartica Linnaeus in the northeast (P. M. Catling and G. Mitrow 2012b).

Hawthorns flower in spring during (rarely before) leaf expansion. They are mass-flowerers such that the anthesis season for a single plant lasts seven to ten days, the timing principally dependent on accumulated heat. Phenological separation is an important evolutionary driver and makes knowledge of flowering sequence very helpful in identification. The length of local flowering season may be as long as six weeks in localities with multiple hawthorns. Pollination is primarily by Diptera and Hymenoptera (N. Power 1980).

Seven species, including Crataegus azarolus Linnaeus, C. mexicana Moçiño & Sessé ex de Candolle, C. pinnatifida Bunge, and C. scabrifolia (Franchet) Rehder as well as those of ser. Aestivales of the flora area, are minor commercial fruit crops. Hawthorn extracts are used to treat hypertension in alternative medicine. In addition to hawthorn, the names haw, thorn, and thorn apple are used colloquially (see J. B. Phipps et al. 2003).

Three ploidy levels (2x, 3x, and 4x) have been demonstrated in a wide range of North American hawthorns both cytologically (various authors) and by flow cytometry (N. Talent and T. A. Dickinson 2005). Diploids are usually sexual outbreeders; triploids are obligate apomicts; tetraploids are self-compatible facultative apomicts; and apomicts are pseudogamous. The polyploid groups are most difficult taxonomically.

Hybridization is routinely touted as a principal cause of taxonomic difficulty in Crataegus. Hybridization within series may be widespread but is hard to detect; the group led by T. A. Dickinson detailed work on it using the C. douglasii complex (for example, E. Y. Y. Lo et al. 2010) is groundbreaking. Putative hybrids between species in different series, often resulting in plants receiving hybrid species rank, are easier to suspect and have been proposed in numerous instances, but on investigation nearly all cases proved non-persistent and only rarely recreated (J. B. Phipps 2005). Series Anomalae, which has some widespread and not uncommon species, may be one exception. Hybrids between species in different sections are nearly unknown today in North America, except for hybrid swarms between the introduced sexual compilospecies C. monogyna and a couple of native diploids (R. Love and M. Feigen 1979; T. C. Wells and Phipps 1989). Ancient, wide hybridization is proposed for the origin of three common, taxonomically isolated, southeastern species (Lo et al. 2009). Hybridization, as contributing to taxonomic difficulty, appears to be restricted to situations involving closely related species or within particular species-complexes.

This treatment of Crataegus is the first attempted for the whole flora area since the work of J. Torrey and A. Gray (1838–1843). It is mainly conservatively based on E. J. Palmer (1950, 1952, 1960), except for series Apricae and Lacrimatae, and new species, mainly western, not known to Palmer.

The taxonomic complexity of North American Crataegus has resulted in extremely narrow species (C. S. Sargent, etc.), middling species (E. J. Palmer), and very broad ones (A. Cronquist). Here, a morphological species concept (MSC) is used. In general, middling positions are favored, and the number of species treated proves similar to what might be extrapolated from work by Palmer except for some southeastern series and western taxa unknown to Palmer. As used here, the MSC assumes that distinguishing suites of character states are adaptive.

A principal objective of this treatment has been to draw attention to the often great, but little discussed, infraspecific variation. This has been possible due to extensive fieldwork by the author and large quantities of available herbarium material. Because of evolutionary implications, infraspecific variation is given considerable attention. Levels of variation in some cases exceed those of woody sexual species of comparable abundance and range yet may not lead to clear dissection into several less variable species. Where high levels of variation occur, varieties are not necessarily recognized, either due to insufficient study or simply because the variation appears chaotic. For example, in the broadly treated Crataegus macrosperma, the infrataxa earlier recognized are not formally treated herein, but detailed discussion of variation is provided. When varieties are used, these themselves may also be variable, as with C. pruinosa and C. viridis. An interesting feature revealed for North American Crataegus is the frequency of occurrence of sympatry in infraspecific variation, this also shown by E. Y. Y. Lo et al. (2009b) in an in-depth study of the C. douglasii complex.

A frequent difficulty for North American Crataegus derives from the co-existence of sympatric, significantly atypical, named entities with otherwise well-defined taxa. They are often uncommon, might be hybrid or simply particularly divergent forms; usually, no one knows. Treatment could be as species (greatly inflating the text), as synonyms (which seems misleading), or dismissed from the flora altogether (which would cripple proper understanding of North American Crataegus variation). Here, they are diagnosed very briefly in the discussion of the most relevant taxon.

Higher level variation is arranged into six sections, these divided into 32 series, mostly following the circumscriptions of E. J. Palmer (1925). Treatment of series may run into similar problems to species delimitation, namely, the existence of atypical forms that obscure taxon limits. Also, some species, for example, Crataegus flabellata (ser. Tenuifoliae), may act as a connecting link to another series, here ser. Rotundifoliae. Both kinds of issue can necessitate repeated keying out of some series. However, rather than amalgamating the series recognized into larger, even more amorphous, ill-defined entities to mitigate these problems, it is preferred to maintain the series in a familiar Palmerian sense preserving the time-tested utility of their core elements.

Species that are similar in flower may often be readily distinguished in fruit and vice versa; users of this text are encouraged to verify identification from the same plant at both seasons.

Variation occurs in the thorns. Compound thorns, which are adventitious and indeterminate, occur on the trunks of many species. Thorns on twigs, which develop with the new growth, persist a few years and are of two kinds. Determinate thorns (aphyllous thorns of K. I. Christensen 1992) are lateral and are the only type found in North American native species. They have diagnostic value, varying in abundance, color in their second year, length, curvature, and stoutness. Indeterminate thorns, as found in Eurasian species of sect. Crataegus, commence as lateral or terminal growths and may develop into thorn-tipped shoot systems.

Leaves supply useful characters for Crataegus taxonomy, but variation within a plant may be confusing. Median leaves of short shoots are the normal standard for descriptions; they are relatively invariant in size and form. Size at anthesis varies according to species from 0% to 100% mature size. Lobing, here often expressed as Leaf Incision Index, or LII, varies from 0% (not incised) to 100% (incised to midvein). As LII typically diminishes towards the leaf apex, maximum LII is usually used. Leaves on extension shoots are often larger, more deeply incised, sometimes of a completely different shape from short-shoot leaves, and may be more variable on one plant. Although apparently diagnostic in some cases, they are insufficiently known and too difficult to describe to be generally useful for description.

Hawthorns may possess both leaflike bracts and bracteoles in the inflorescence, the former being either relatively few or absent. Bracteoles are often relatively numerous and can usually be found throughout young inflorescences; they are sparse in some species. Typical bracteoles of North American hawthorns are caducous, more or less linear-oblong to linear, bilaterally symmetric, and membranous with glandular margins. Some North American hawthorns have larger, harder, more or less herbaceous and more persistent bracteoles. A different bracteole type found in some North American hawthorns (for example, Crataegus brachyacantha) is smaller than any of the above and more or less lacking marginal glands.

Fruit color change during development is a valuable distinguishing character in some cases. In comparing two sympatric species, it is critical to compare them at the same time and a good example is found with Crataegus okennonii versus C. douglasii. Although pomes both finish black, pomes of C. douglasii are usually already black when those of C. okennonii are brownish or reddish, as illustrated in J. B. Phipps et al. (2003).

Crataegus is best treated as sister to Mespilus (E. Y. Y. Lo and M. Donoghue 2012; Li Q. Y. et al. 2012); other relationships in tribe Pyreae are less close. Molecular evidence (C. S. Campbell et al. 2007; D. Potter et al. 2007) shows that the Amelanchier clade is sister to Crataegus-Mespilus, but Amelanchier is different in many morphologic features.

The following names could not be accounted for in the current treatment: Crataegus ardua Sargent, C. brachyphylla Sargent, C. evansiana Sargent, C. faxonii Sargent var. praetermissa (Sargent) E. J. Palmer, C. glareosa Ashe, C. harryi Sargent, C. immanis Ashe, C. lecta Sargent, C. littoralis Sargent, C. menandiana Sargent, C. mercerensis Sargent, C. pilosa Sargent, C. pinguis Sargent, C. populifolia Walter, C. sublobulata Sargent, and C. whitakeri Sargent.

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

Genera 88, species ca. 3000 (68 genera, 680 species, including 22 hybrids, in the flora).

Three subfamilies and 16 tribes are recognized for the family with representatives of all tribes found in the flora area.

Rosaceae grow most commonly in north-temperate regions and are more or less absent from hot deserts and high-rainfall, low-altitude tropics. The family is large and diverse, characterized by radially symmetric flowers with a fundamentally saucer-shaped hypanthium and peripheral calyx, corolla, androecium, and, usually, superior gynoecium. Considerable variation occurs in important details of flower and fruit, this enhanced by a very adaptable hypanthium, which is discussed in more detail below.

Rosaceous inflorescences vary in the number of flowers from one to about 500. A great variation occurs also in inflorescence form, though a pattern is perceived when they are viewed as reduction series from a terminal, determinate panicle that is fundamentally bracteate, thus generating a range of panicles, racemes, corymbs, cymes, solitary flowers, or other forms. In this treatment, appendages on the inflorescence are distinguished by the terms bract and bracteole. Bract is used for the larger, laminate, usually chlorophyllous leaf homologues that subtend axes and may be indistinguishable from foliage leaves. Bracteoles are scalelike, often membranous, often caducous, and of uncertain homology, in part due to not being restricted to axis-subtending positions.

The floral architecture in Rosaceae is radially symmetric around a disc-shaped to urceolate hypanthium. Flowers normally have a four- or five-merous corolla and calyx; great variation is found in the numbers of stamens and carpels. Pollination is usually entomophilous, the flowers having normally green sepals and showy, often more or less clawed, usually white, yellow, or pink, less commonly red or green, petals. The flowers in some genera are relatively small and anemophilous and may lack one or two of the principal whorls. An unusual feature of some rosaceous flowers is the torus, a pad of receptacular, usually spongy, tissue, sometimes relatively large, in the center of the hypanthium that, when present, bears the gynoecium. Another unusual feature of some genera is an epicalyx that comprises a ring of sepaloid bractlets, usually of the same number as sepals, which is located on the hypanthium proximal to the calyx.

Fruit types are particularly significant both in rosaceous identification, taxonomy, and diversity, as well as for successful dispersal. Fruit in Rosaceae may be either dry, then dehiscent or not, or succulent and indehiscent; it sometimes involves accessory organs such as the torus or a persistent hypanthium.

Dry indehiscent fruits are achenaceous, in certain cases involving modifications to the style. In anemochorous (adapted for dispersal by wind) situations, this may involve a plumose style, for example, Cercocarpus; in epizoochorous (distributed on the outside of animals) cases, the styles may bear stiff hairs or barbs, for example, Geum. Alternatively, achenaceous fruits may lack styles (sometimes due to abscission) or, more commonly, have a relatively short one, for example, Potentilla. In situations where there appears to be no significant post-flowering function for the style, dispersal may be myrmecochorous (by ants), anemochorous, or in the rare torus-borne cases (for example, Fragaria), the fruit may be endozoochorous (eaten by animals and passed through the gut). Sometimes, fruits with such styles are aggregated in acheneta that rely on barbed, persistent hypanthia for epizoochorous dispersion, for example, Acaena and Agrimonia. Dry dehiscent fruits are follicular with seeds normally distributed by air after splitting of the ripe follicle, for example, Gillenia.

Succulent, endozoochorous fruits exhibit a similar range of variation. The most common types are: multiple drupelets on the surface of a more or less conic torus (for example, Rubus); individual and sometimes large drupes on a flat receptacular apex (Prunus); pomes, which are berrylike fruits in which a fleshy hypanthium more or less completely surrounds and is generally more or less fully adnate to the carpels (for example, Amelanchier, Crataegus). The result is that a terminal orifice may remain open, often bearing floral remnants around its rim. Pomes with hard pyrenes are sometimes distinguished as polypyrenous drupes, a term not used in this treatment due to antonymic confusion caused by combining the roots for pyrene (hardened mesocarp) and drupe (fleshy mesocarp), although the relevant distinction is well recognized (J. Rohrer et al. 1991).

Some important temperate fruits are members of the Rosaceae: apples (Malus), pears (Pyrus), almonds, apricots, cherries, peaches, and plums (Prunus), blackberries and raspberries (Rubus), strawberries (Fragaria), loquat (Eriobotrya); minor fruits include those of Amelanchier, Crataegus, Cydonia, Mespilus, and others. Some genera are popular in ornamental horticulture in North America, for example, most of the above as well as, especially, Chaenomeles, Cotoneaster, Pyracantha, Rhaphiolepis Lindley, Sorbus, Photinia, Physocarpus, Rosa, and Spiraea among trees and shrubs; and Filipendula, Geum, Potentilla, and Spiraea among plants suitable for flower borders.

Apomixis is a feature of some rosaceous genera and may make taxonomic decisions difficult or equivocal in genera such as Alchemilla, Amelanchier, Crataegus, Rubus, and others. Apomixis always seems to be associated with polyploidy and often with hybridization; in some apomictic genera, sexual species and apomicts are facultatively sexual.

Rosaceae lack alkaloids and blue anthocyanic pigments. Some have foliage or seed that becomes toxic due to hydrolysis of benzaldehyde cyanohydrins.

The great diversity of Rosaceae reflects the age of the family (since at least the mid Eocene) and its evolutionary success. The family has needed no fundamental change in circumscription since Jussieu first recognized it, the sometimes included outgroups, for example, Chrysobalanaceae, Neuradaceae, already being confidently excluded in more recent pre-molecular classifications, while segregate families (for example, Malaceae) have received little currency. This reflects a lack both of close neighbors and of large internal discontinuities. Species limits are often debatable, especially where apomixis is present; generic limits are much more fixed, apart from the familiar historic trend to recognizing smaller, more discrete units. What has principally been fluid in rosaceous taxonomy until modern molecular research are the phylogenetic relationships, both to neighboring families and within the family. Rosales once contained families such as Crassulaceae and Saxifragaceae, which have many morphological similarities to Rosaceae, but Angiosperm Phylogeny Group (2003) indicated instead that in Rosales, Rosaceae is sister to a group of families including Moraceae, Rhamnaceae, Ulmaceae, and Urticaceae.

The internal relationships of Rosaceae have received much study, including morphological, anatomical, cytological, phytochemical, breeding system, fungal pathogenicity (especially rusts), and molecular. Among these studies, that of D. Potter et al. (2007) resulted in the first comprehensive molecular phylogeny of Rosaceae, and the treatment in this volume reflects it. Potter et al. recognized three subfamilies with 15 tribes worldwide. Their classification used the ranks supertribe and subtribe, not used here. The subfam. Dryadoideae, with one tribe and four genera, diverged early and is unique in Rosaceae for its actinorhizal symbiosis; it has achenaceous fruits. Subfamily Rosoideae has six tribes and 29 genera and is only slightly altered from its traditional circumscription by shedding Dryadoideae and three small genera. Most Rosoideae have achenaceous fruit; some have multiple drupelets or achenes borne on a torus. The remainders of Rosaceae are found in the large and heterogeneous subfam. Amygdaloideae of nine tribes and 54 genera. This subfamily has the largest diversity of fruit types. Amygdaloideae now contains the traditional spiraeoid genera with follicular fruit, Maleae with pome fruit, and Amygdaleae with drupaceous fruit.

Key to Subfamilies and Tribes of Rosaceae

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