Salix

Shrubs or trees, slightly heterophyllous, clonal or not, clones formed by root shoots, rhizomes, layering, or stem fragmentation; branching sympodial. Stems not spinose. Buds 1-scaled (oily in S. barrattiana), margins connate into calyptra or distinct and overlapping adaxially, scale inner membranaceous layer usually not separating from outer layer, (sometimes free and separating). Leaves deciduous or marcescent; stipules persistent, caducous, or absent (varying in presence and size on early and late leaves); petiole glandular-dotted or lobed distally; (blade often more than twice as long as wide, venation usually pinnate, margins entire, crenulate, crenate, serrate, serrulate, or spinulose-serrulate, teeth gland-tipped). Inflorescences axillary or subterminal, catkins, erect, spreading, or ± pendulous, sessile or terminating flowering branchlets, unbranched (except in subg. Longifoliae); floral bract apex entire, erose, 2-fid, or irregularly toothed; pistillate bract persistent or deciduous after flowering. Pedicels present or absent. Flowers: (sessile), perianth reduced to adaxial nectary (rarely also abaxial nectary, then distinct or connate into shallow cup); stamens 1, 2, or 3–10; filaments distinct or connate; ovary (stipitate or sessile), 2-carpellate; ovules (2–)4–24(–42) per ovary; styles usually connate, sometimes distinct distally; stigmas 2, entire or 2-lobed (less than 2 mm). Fruits capsular, (2-valved, obclavate to ovoid or ellipsoid). Seeds: aril present. x = 19.

Distribution

North America, Mexico, West Indies, Central America, South America, Europe, Asia (Malaysia), Africa, Atlantic Islands, mostly in arctic, boreal, and temperate regions, introduced in Australasia, Oceania.

Discussion

Species ca. 450 (113 in the flora).

Species of Salix have been studied by taxonomists, morphologists, anatomists, geneticists, cytologists, chemists, ecologists, arborists, entomologists, and others. Classification of the genus and identification of specimens remains difficult. C. K. Schneider (1919b) stated, “In determining willows one is only too often entirely misled at first, [but] even by a slow and careful examination it is not always possible to determine the proper identity of the plant.”

Classification. Traditionally, the subgeneric classification of Salix was based on morphological characteristics; recent molecular studies have begun to provide useful insights. The first classification of New World Salix (C. K. Schneider 1921) recognized 23 sections and arranged them in linear order corresponding to usually recognized subgenera. The first classification to use subgenera (R. D. Dorn 1976) recognized two: subg. Salix (including tree willows and sect. Longifoliae) and subg. Vetrix (including shrubby and dwarf arctic-alpine willows). G. W. Argus (1997) recognized four subgenera: Chamaetia, Longifoliae, Salix, and Vetrix. In the present classification, five subgenera are recognized, with subg. Salix being divided into subg. Protitea (bud-scales with distinct, overlapping margins and flowers with multiple stamens) and subg. Salix (bud-scales with connate margins and flowers usually with two stamens; see discussion under 2a. subg. Protitea).

There are two published studies of Salix classification based on molecular data (see discussion under subgenera Protitea and Longifoliae). In both studies, the number of Salix species included is relatively low. E. Leskinen and C. Alström-Rapaport (1999) studied phylogeny of Salicaceae and Flacourtiaceae and sought to determine the relationship of Chosenia to Salix. Parsimony analysis showed little resolution within Salix and bootstrap support was strong for only three relatively small groups. Salix and Chosenia were placed in a single clade with two major branches: 1) the first included S. exigua (subg. Longifoliae); 2) the second included two subgroups: 2a) S. amygdaloides (subg. Protitea) and S. alba, S. euxina, and S. pentandra (subg. Salix); 2b) all the other species (subg. Chamaetia and Vetrix). The species in this subgroup were unresolved. T. Azuma et al. (2000) sought to determine the taxonomic position of Chosenia and Toisusu, as well as the classification of Salix. Within Salix two major clades were recognized. Clade 1 consisted of three major branches: 1a) included S. interior (subg. Longifoliae) along with S. amygdaloides and S. nigra (placed here in subg. Protitea); 1b) included Asian S. chaenomelioides (subg. Pleuradinea); and 1c) included African and Asian S. safsaf and S. tetrasperma (subg. Protitea). Clade 2 also had three major branches: 2a) S. triandra (as S. subfragilis); 2b) unresolved members of subg. Chamaetia and Vetrix; and 2c) Asian genera Toisusu and Chosenia (now placed in subg. Pleuradinea). Molecular data, although not conclusive because of unresolved species, lend support for recognition of subg. Longifoliae and subg. Protitea. Further studies are needed to refine the subgeneric classification.

Biology. Salix are pioneer or early succession species well-adapted to disturbance. Each pistillate plant can produce hundreds to thousands of seeds annually. At maturity, the seeds can be lifted into the air surrounded by a parachute of fine hairs that can carry them tens to hundreds of meters from the parent plant. Seeds that land on water can float for several days because of the flotation capability of the hairy hilar aril surrounding each seed (E. M. A. Steyn et al. 2004). Willow seeds have no food reserves; most will perish within days unless they land and germinate in a suitable habitat. Some arctic and subarctic species (R. A. Densmore and J. C. Zasada 1983) and some members of sect. Salicaster (see 12. S. serissima) are able to survive through the winter and germinate in the spring. Some tropical species flower year-round (P. Parolin et al. 2002) by producing sylleptic catkins and, thus, are able to take advantage of newly disturbed habitats at any time.

Seedling success depends mainly on an adequate supply of moisture and the absence of shading (C. F. Sacchi and P. W. Price 1992). When such habitat is available, seeds will germinate almost immediately upon arrival. During the Pleistocene, such conditions were present on a large scale because repeated glacial and interglacial periods provided suitable environments for some northern Salix to acquire their present circumpolar or transcontinental distributions. Willow seedlings most commonly occur on riparian sand and gravel bars, old burns, landslides, drained lakes and wetlands, and in open, unstable arctic and alpine habitats. They also are common along the vast network of roads that crisscross even some of the most remote wilderness areas. Road margins, ditches, and gravel and sand borrow pits provide favored habitats. Even minor disturbances, such as upturned tree roots, ungulate tracks in wet meadows and mires, cryogenic frost boils and cracks in tundra, and animal diggings, can provide willow habitat. Because of this need for open habitats for reproduction by seed, large stands of mature willows growing in stable habitats such as marshes, fens, bogs, treed riverbanks, and even active sand dunes, have become established in these habitats before a closed cover had developed. While individual plants can sometimes invade closed vegetation, large stands of willows require large disturbances.

The zonation of willows on floodplains is a function of seed dispersal timing and water level fluctuations (L. R. Walker et al. 1986; I. Van Splunder et al. 1995). Seedling success depends on abiotic factors, such as erosion by flooding later in the season and siltation in subsequent years, and on biotic factors, mainly herbivory by moose and snowshoe hares. The colonization of glacial moraines of different ages (G. W. Argus 1973) showed that all species of Salix in the area colonized the earliest moraines. Over time, the dwarf and presumably less shade-tolerant species, S. arctica, S. reticulata, and S. stolonifera, were eliminated, followed by the low to mid shrubs until only tall shrubs and trees of S. sitchensis remained along streams and in openings in the spruce-fir forests.

Some willows are adapted for vegetative reproduction by stem fragmentation, layering, or root shoots, and all species can collar-sprout at or below ground level (P. Del Tredici 2001). Vegetative reproduction by stem fragmentation is characteristic of riparian species, some of which have brittle branches, which can be dispersed by wind and water to where they may become lodged and root. Species that spread by root shoots or layering are more limited in their dispersal potential but often form distinctive clones.

Most willows can be propagated by cuttings; some root more easily than others. Adventitious root primordia are initiated mainly at the base of cuttings, sometimes at or between nodes, under stimulus of auxin or other compounds that migrate to the basal end (B. E. Haissig 1974). Riparian species (e.g., Salix alaxensis, S. lasiandra, S. pseudomyrsinites) usually have preformed primordia and root along the entire cutting; non-riparian species (e.g., S. bebbiana, S. glauca, and S. scouleriana), usually described as rooting poorly, do not have preformed primordia (R. A. Densmore and J. C. Zasada 1978). This led Densmore and Zasada to suggest that preformed root primordia are an adaptation to the flooding and siltation that occur in riparian habitats. The formation of root primordia is more complex than that. Willows of non-riparian habitats (bogs, fens, prairies, sand dunes, tundra, and upland forests) display diverse rooting patterns. All of these habitats may have aggrading surfaces due to siltation, moss and peat accumulation, sand drifting, etc., that can encourage the formation of adventitious roots. Bog and fen species (S. candida, S. fuscescens, and S. pedicellaris) typically root along their stems, as do plants growing in sand dunes (S. brachycarpa, S. cordata, and S. silicicola). Even upland species (S. bebbiana, S. humilis, and S. scouleriana), usually regarded to root poorly, can root prolifically when growing in wet or aggrading habitats.

Morphology. Salix are woody plants varying from trees reaching 30 m, to dwarf arctic-alpine shrubs less than 5 mm. They often form clones by stem fragmentation, layering, rhizomes, or root shoots. Branchlets are current-year stems, and branches are stems more than one-year old. Buds have a single scale. The bud-scale margins usually are connate; in subg. Protitea they are distinct and overlapping. Because shoot growth in Salix is sympodial, buds at shoot apices are subterminal. Vegetative and reproductive buds vary in size, shape, and position. Three general types of bud size and shape gradation are recognized; namely, alba-type, arctica-type, and caprea-type; there are intermediates (A. K. Skvortsov 1999). Plants with alba-type bud gradation have buds that are very similar in size and shape along branchlets (monomorphic), but floral and vegetative buds cannot be distinguished from one another. Plants with arctica-type bud gradation usually have relatively few buds. The two or three (sometimes to five) distal buds are the largest, diminishing in size proximally. Usually only the larger buds open and those buds may be either floral or vegetative. Plants with caprea-type bud gradation have floral buds that are strikingly different in size and shape from vegetative buds (dimorphic). Usually, the distal two or three buds are vegetative, the next three to six (or more) are floral, and proximal to them are smaller vegetative buds.

Stipules borne on either side of the petiole may be foliaceous, minute rudiments, or absent. Those on early (preformed) leaves often are rudimentary; those on late (neoformed) leaves often are foliaceous. Although not all leaves are differentiated in the winter buds (E. Moore 1909), it is difficult to determine precisely which leaves are neoformed. While it is probable that morphological differences occur among leaves on an individual plant, as in some Populus species (W. B. Critchfield 1960), only stipule differences have been noted. Petioles sometimes have glandular-spherical dots or lobes at the distal end just proximal to the blade. Sometimes the petioles are ventricose or inflated around the subtended floral buds.

Three types of leaves are recognized in Salix: large medial blades are the “normal” leaves; proximal blades are the first two to four reduced true-leaves at the base of branchlets or on catkin-bearing shoots and differ from distal (late) ones in shape, indumentum, dentition, and prominence of stipules; and juvenile blades are young unfolding leaves at the distal end of branchlets. These leaves vary in shape from linear to subcircular, bases are cuneate to cordate, margins are entire or crenate to spinulose-serrulate, and leaf teeth are gland-tipped. Glands on teeth, or entire margins, may be marginal, submarginal (blade edge viewed from abaxial surface has a margin slightly revolute or thickened), or well up on the adaxial surfaces (epilaminal). Abaxial blade surfaces, and sometimes adaxial, are often glaucous with a dull, waxy coating; blade surfaces may be glabrous or hairy, leaf hairs (trichomes) are usually white, sometimes ferruginous (rust-colored). Syllepsis, the opening of buds without a rest period, is common in subg. Longifoliae, as well as some Populus, and has been recorded in 19 species of Salix representing all subgenera. Sylleptic leaf morphology sometimes differs from proleptic leaves (see 2c. subg. Longifoliae).

The inflorescences are catkins (aments), each of which consists of a flower-bearing rachis (essentially a spike of unisexual, apetalous, sessile flowers, each subtended by a floral bract), and a peduncle. A catkin may be sessile on a branch or borne on a relatively short, vegetative flowering branchlet (a shoot bearing three or more green leaves). Catkins arise from lateral or subterminal buds. They flower before leaves emerge (precocious), as leaves emerge (coetaneous), or throughout the season. For use of the term serotinous, see 12. Salix serissima. The flowering rachis is usually unbranched; in subg. Longifoliae secondary or tertiary branching can occur. After anthesis, the rachis of a pistillate catkin continues to elongate but rachises of staminate catkins do not. A floral bract (scale) subtends each flower. Pistillate floral bracts are usually persistent in fruit; in some subgenera they are deciduous. Each flower consists of an adaxial nectary (a reduced perianth, according to M. J. Fisher 1928) located between the stamens or pistil and rachis axis; in some taxa there is also an abaxial nectary located between the floral bract and fertile structures; the two nectaries may be distinct or connate into a cup-like structure. Staminate flowers usually have two stamens, but the number can be one or three to ten. Pistillate flowers have a single pistil, which is sessile or borne on a stipe (Fisher; S. Sugaya 1960). There are two styles, usually connate, each terminated by a two-branched stigma. Stigma lobes are: 1) flat abaxially, papillate adaxially and with a rounded or pointed tip; 2) slenderly cylindrical (length greater than four times the width) or broadly cylindrical (length less than four times the width); or 3) subspherical (plump). The number of ovules per ovary can be determined by counting funiculi remaining in mature capsules after seeds have been shed. For definitions of Salix terminology, see G. W. Argus (2007).

Variability. Some species of Salix are highly variable and closely related species may be only subtly distinct. Underlying most identification problems is morphological variability, some of which is related to biology of the genus. Some phenotypic variability can result from habitat modification. For example, shade conditions can reduce the density of leaf and branch glaucescence (R. D. Dorn 2003), as well as leaf thickness. The most important sources of morphological variation are hybridization, introgression, and allopolyploidy.

In describing the ‘gloss’ of stems and leaves in Salix, three classes are used: dull, slightly glossy, and highly glossy. These three classes intergrade, but, usually, they can be separated. The modifier ‘highly’ is used with the word glossy to emphasize that this condition is a distinct extreme of glossy, even though in many species there is gradation from one to the other. It is best to understand the distinctions by example; the adaxial leaf surface of S. petiolaris is glossy or dull, whereas in S. serissima it is highly glossy (analogous to a varnished surface or freshly waxed floor). Only a few species of Salix (S. caroliniana, S. floridana, S. maccalliana, S. nummularia, S. ovalifolia, S. pentandra, S. phlebophylla, S. planifolia, S. rotundifolia, S. serissima, S. stolonifera, S. tyrrellii) have leaves that are described as only highly glossy. In these species, the ‘gloss’ of the adaxial leaf surface is a diagnostic character.

Hybridization. Approximately 120 Salix hybrids have been recognized in the North American flora, and about half of these are relatively common. Others are either putative hybrids in which one parent may be uncertain or unconfirmed, and/or they are doubtful hybrids. North American botanists, in general, have been conservative in their recognition of hybrids, probably in reaction to some European botanists who readily recognized not just simple hybrids but multiple-species hybrids. In Greenland, B. G. O. Floderus (1923) recognized five pure Salix species and seven interspecific hybrids, some of which were three-species hybrids. Working in a similar flora, H. M. Raup (1943, 1959) argued against an uncritical recognition of hybrids and suggested that intermediate specimens should be given the name of the species they resemble most. These views were shared by A. K. Skvortsov (1999), who agreed that willows are inherently variable and that a better understanding of species variability would reduce the number of presumed hybrids.

There are barriers to hybridization, including differences in flowering time (A. Mosseler and C. S. Papadopol 1989), pollen-stigma incompatibility (Mosseler 1989), and F1 hybrid inviability. Nevertheless, hybridization among Salix species can be an important source of variability. Hybridization, clonal reproduction, and the ability of hybrids to backcross may be accompanied by introgression (J. Salick and E. Pfeffer 1999). Hybrids can sometimes be recognized by discordant character variations, such as the occurrence of partially hairy ovaries within species characterized by glabrous ovaries, or by having leaf surfaces glaucous abaxially within species that characteristically lack leaf glaucescence. Sometimes such variation may occur along with teratological flowers, or other evidence of infertility and reproductive imbalance. Hybrids, which can be difficult to recognize in the herbarium, sometimes are recognizable in the field as being different from other individuals nearby.

A morphological and molecular study of hybridization and introgression between Salix eriocephala and S. sericea (T. M. Hardig et al. 2000) found that ca. one-third of plants originally identified as S. eriocephala were possible introgressants. Other plants showed unequivocal evidence of backcrossing with S. sericea; inter- and intra-specific chloroplast diversity found within a hybrid zone suggested both historic introgression, perhaps in a glacial refugium, and contemporary hybridization. Hardig et al. found that hybrids might not be readily recognized in either the field or herbarium and wrote that, “If major distinguishing characters are under the control of one or two dominant genes, hybridization may go unrecognized. Important taxonomic characters that are quantitative might result in recognizably intermediate hybrids but … hybrids may be imperfectly intermediate or highly variable, resulting in an interpretation that unrecognized hybrids are merely part of the morphological variation in one of the species.” The practical taxonomic message is that the interpretation of species variation as either inherent or due to hybridization must be made carefully. While it may be unwise to mistake hybridization for species variability, it is equally unwise to mistake species variability for hybridization.

Polyploidy. The widespread occurrence of polyploidy in Salix is an important indication of the evolutionary importance of hybridization. Among the 99 native Salix species in the flora area, 70% have chromosome counts and, of those, 47% are polyploid. It can be assumed that 50% or more of the native Salix are polyploid. It is probable that most of these are allopolyploids, inasmuch as there is little evidence of autoploidy in Salix (W. Buechler, pers. comm.). Some of the most variable species may have evolved through hybridization and polyploidy. For example, S. arctica and S. glauca each display several ploidy levels, as well as many hybrids; both species probably evolved through repeated hybridization and backcrossing with each other and other species. The possibility that recurrent polyploidy has contributed to variability in these and other polyploid species (R. J. Abbott and C. Brochmann 2003) needs study.

Collection and identification. Because Salix are dioecious, a single individual cannot provide the full range of reproductive and vegetative structures needed for identification. Some species flower well-before leaves emerge; reproductive structures, especially staminate catkins, and foliage may not be available simultaneously. Ideal specimens for identification include flowering, fruiting, vegetative, and winter twigs. It is possible to gain an in-depth understanding of seasonal morphological variability by tagging plants and making collections at different developmental stages. Insights into population variability, hybridization, and introgression can be gained by attempting to identify every plant in a stand (A. K. Skvortsov 1999). At a minimum, well-collected and pressed specimens are essential. All available plant parts should be collected: leaves (including juvenile leaves), catkins, and twigs. Sprouts, or compensatory shoots, are not commonly included in keys or descriptions. If collected at all, they should be to supplement normal shoots and labeled as such. To avoid loss of glaucescence (wax on stems or leaves), specimens should be dried as rapidly as possible but without using excessive heat. Plant habit and evidence of vegetative reproduction should be noted.

Keys. Variability makes writing dichotomous keys to Salix taxa difficult. At best, the keys often are cumbersome to use and may only account for a relatively small part of variability within a taxon. The preparation of separate keys to staminate, pistillate, and vegetative specimens is useful but such keys are not usually provided (G. W. Argus 1986; A. Cronquist and R. D. Dorn 2005). One of the best ways to identify specimens is to use interactive keys (M. J. Dallwitz et al., http://delta-intkey.com; R. J. Pankhurst 1991). An interactive key to New World Salix using Intkey is available (G. W. Argus, http://aknhp.uaa.alaska.edu/willow). It can be used not only to identify specimens, but to describe or compare species, or to list species by state, province, or taxonomic group.

Uses. Willows play major roles in ecosystems by rehabilitating disturbed sites through stabilization to prevent erosion, to improve soil, to remove pollutants and heavy metals, and to provide wildlife food and habitat. Willows are used widely as ornamentals. Most introduced species are ornamental cultivars. In many parts of the world, willows are used in basketry, as sources of tannins, and in apiaries as food for brood rearing and making honey. Traditionally, willows were used in medicines; salicin (a component of aspirin) was first derived from Salix. Their use as a source of energy biomass is being investigated worldwide. Indigenous peoples have used willows for fuel, construction, basketry, medicines, tools and weapons, and ceremonially.

Conservation. Inasmuch as willows are pioneer species, they present special conservation problems. Attempts to protect them by preventing habitat disturbance will be counterproductive. Although they do not spread without disturbance, once established they may require protection against biotic factors, such as browsing, so that they can produce propagules and disperse them to nearby disturbed sites (see 59. Salix arizonica). Given the opportunity, some non-native willows can be very aggressive (see 81. S. cinerea). In Australia and New Zealand, where Salix is not native, some introductions have been so successful that they are regarded as invasive weeds (C. J. West 1994), and control measures are being implemented. In the flora area, some introductions, such as S. ×fragilis, have spread so readily by stem fragmentation that, although they rarely produce seed, they appear to be part of the native flora.

Notes on Style. In the descriptions of taxa and in the keys, quantitative morphological data have been given in three ways, depending on degree of variation and sample size. For example: 10–100 [minimum–maximum]; (3–)10–75(–100) [(rare extreme–)usual minimum–usual maximum(–rare extreme)]; 10–20–100 [minimum–mean–maximum]. For Salix distribution maps with more detailed ranges, see G. W. Argus (2007).

Selected References

Argus, G. W. 1973. The Genus Salix in Alaska and the Yukon. Ottawa. [Natl. Mus. Nat. Sci. Publ. Bot. 2.]

Argus, G. W. 1986. The genus Salix in the southeastern United States. Syst. Bot. Monogr. 9.

Argus, G. W. 1986b. Studies in the Salix lucida Muhl. and S. reticulata L. complexes in North America. Canad. J. Bot. 64: 541–551.

Argus, G. W. 1995. Arizona Salicaceae: Willow family. Part two: Salix. J. Arizona-Nevada Acad. Sci. 29: 39–62.

Argus, G. W. 1997. Infrageneric classification of Salix L. in the New World. Syst. Bot. Monogr. 52.

Argus, G. W. 2007. Salix L. (Salicaceae): Distribution maps and a synopsis of their classification in North America, north of Mexico. Harvard Pap. Bot. 12: 335–368.

Azuma, T., T. Kajita, J. Yokoyama, and H. Ohashi. 2000. Phylogenetic relationships of Salix (Salicaceae) based on rbcL sequence data. Amer. J. Bot. 87: 67–75.

Dorn, R. D. 1976. A synopsis of American Salix. Canad. J. Bot. 54: 2769–2789.

Dorn, R. D. 1977b. Willows of the Rocky Mountain states. Rhodora 79: 390–429.

Dorn, R. D. 1997. Rocky Mountain Region Willow Identification Field Guide. Denver.

Raup, H. M. 1943. The willows of the Hudson Bay Region and the Labrador Peninsula. Sargentia 4: 81–135.

Raup, H. M. 1959. The willows of boreal western America. Contr. Gray Herb. 185: 1–95.

Lower Taxa

Salix subg. Chamaetia, Salix subg. Longifoliae, Salix subg. Protitea, Salix subg. Salix, Salix subg. Vetrix

Key

Key to Subgenera of Salix

1	Bud-scale margins distinct, overlapping; stamens 3-7[-9]; pistillate bracts deciduous after flowering (except S. floridana, sometimes S. bonplandiana); usually trees.	Salix subg. Protitea
1	Bud-scale margins connate; stamens usually 2, sometimes 1 (3-10 in sect. Salicaster); pistillate bracts persistent (except in sect. Salicaster and in subg. Longifoliae); shrubs or trees	> 2
2	Stamens 3-10; pistillate bracts deciduous after flowering.	Salix subg. Salix (sect. Salicaster, sect. Triandrae)
2	Stamens 2 or 1; pistillate bracts persistent after flowering (except in subg. Longifoliae, S. alba, and S. euxina)	> 3
3	Pistillate bracts deciduous after flowering; leaf blades usually linear to narrowly elliptic; catkins sometimes branched; syllepsis common; plants clonal by root shoots; petioles not glandular distally.	Salix subg. Longifoliae
3	Pistillate bracts persistent after flowering (except in S. alba and S. euxina); leaf blades usually not linear, rarely narrowly elliptic; catkins not branched; syllepsis uncommon; plants not clonal by root shoots (except S. setchelliana); petioles glandular or not distally	> 4
4	Petioles usually glandular-lobed or -dotted distally; branches and branchlets brittle at base.	Salix subg. Salix
4	Petioles not glandular distally or glands simple, spherical; branches flexible at base (sometimes brittle in subg. Vetrix)	> 5
5	Buds arctica-type or transitional or alba-type; catkins arising from subterminal buds (sects. Chamaetia, Herbella, Myrtosalix) as well as lateral buds; shrubs 0.005-6 m; juvenile blade hairs usually white, rarely ferruginous (in S. athabascensis); pistillate catkins always on flowering branchlets; largest medial blades (relatively broad) 0.8-5.5 times as long as wide; ovaries sometimes glaucous (sects. Chamaetia, Diplodictyae, Myrtilloides, Ovalifoliae), hairs sometimes flat and ribbonlike (sect. Myrtosalix); most staminate and some pistillate flowers with abaxial and adaxial nectaries; inner membranaceous bud-scale layer not separating from outer layer.	Salix subg. Chamaetia
5	Buds alba-type or intermediate; catkins arising from lateral buds (subterminal buds vegetative); shrubs or trees, 0.1-20 m; juvenile blade hairs mostly white, sometimes ferruginous; pistillate catkins sessile or borne on flowering branchlets; largest medial blades (relatively narrower) 0.7-13.7 times as long as wide; ovaries not glaucous, hairs not flat and ribbonlike; staminate and pistillate flowers usually without abaxial nectaries; inner membranaceous bud-scale layer sometimes distinct and separating from outer layer.	Salix subg. Vetrix