There are no active references in this article.
The genus Salix comprises approximately 400 species and has an almost worldwide distribution (it is absent from Australia and Antarctica), though there is a greater diversity of species in the northern temperate zone. Salix species are dioecious woody plants; most are erect trees or shrubs, often with pendulous branches, although a significant minority are prostrate shrubs that occur beyond the treeline in the Arctic or in alpine areas. Terminal buds are usually absent; winter buds with one scale. The leaves are deciduous, alternate or rarely opposite, and remarkably variable. They range in size from 25 cm long (S. magnifica) to only 1 cm long (S. serpyllifolia); they may be thin or leathery, hairy or glabrous, with shapes ranging from lanceolate to circular. Considerable leaf variation occurs within species, and leaves of juvenile shoots are often completely different from those of mature branches. Hybridisation, which is extremely common in Salix, further extends the range of leaf variation. Stipules are present in most species. Inflorescences (catkins) are erect and are produced either before the leaves emerge (precocious), during leaf emergence (coetaneous) or afterwards (serotinous); floral bracts are entire. The flowers have no perianth, and one to two nectaries. The fruit is a two-valved capsule containing numerous seeds surrounded by hairs (Argus 1986, Newsholme 1992, Fang et al. 1999). Phylogenetic studies based on DNA have demonstrated that the genera Chosenia and Toisusu should be included in Salix (Azuma et al. 2000b).
The vast array of willows provides trees and shrubs for all areas of the garden and arboretum, and some species are used in the production of a range of commercial items, from cricket bats to biomass fuel. Much horticultural interest focuses on those with coloured stems in winter, usually grown as coppice or pollards, or those with handsome male catkins in early spring, again often pruned to achieve vigorous new shoots with abundant inflorescences. As larger trees, willows are usually relegated to the damper edges of the arboretum, although they do not necessarily need particularly damp conditions. There is good coverage of Salix in Bean (1981b), but essential reading is Christopher Newsholme’s Willows: The Genus Salix (1992). Propagation is either by seed, sown as soon as it ripens, or by hardwood cuttings in winter. Although these are notoriously easy to root, best results are achieved by taking cuttings in autumn to enable good early callus formation; when growth resumes in spring, shoot and root growth is then in balance.
Salix is a large genus, though the number of species in it depends very much on how narrowly or broadly these are defined. By some reckonings it has 500 species, by others only 300. They vary from stately timber trees such as S. alba or S. nigra, to tiny shrubs like S. herbacea, creeping along the ground and only rising an inch or two above it. But the great majority of the species are erect shrubs or small trees. The twigs of many species are very tough and flexible, and supply much the greater part of the material from which baskets and wickerwork are made in the temperate parts of the northern hemisphere. But the twigs of some, though tough, are easily snapped off in their entirety at the point of union with the older branchlet. These root readily in a favourable soil. This curious characteristic is best known in, and gives the popular name to, the crack willow, S. fragilis, but there are several more that have it equally marked. The winter-buds of the willows are covered with a single scale; terminal winter-buds are not formed (except in a few dwarf alpine willows).
The leaves are deciduous (semi-persistent in a few mostly subtropical species), simple, usually short-stalked, toothed or entire, alternate or, much more rarely, nearly or quite opposite, notably in S. purpurea and some of its allies. They are very variable in shape, usually lanceolate and slender-pointed in tree-willows, sometimes almost orbicular in dwarf species, and with every gradation in between. In the remarkable S. magnifica from China, the leaves are over 8 in. long and 5 in. wide; at the other extreme, the alpine shrublet S. serpyllifolia has them less than 1⁄2 in. long. Stipules produced by willows are variable both in their size and persistence. They occur most markedly on the strongest shoots, and are frequently entirely absent from weaker ones. In some tree species glands are present near the junction between leaf-blade and petiole, and are sometimes converted into minute stipules.
The inflorescence is a catkin (ament) borne at the tip of a shoot that (in most species) springs from an axillary bud on the previous year’s wood. This flowering shoot may be 2 in. or even more long, and bear leaves that are normal except in having no growth-buds in their axils, the whole shoot being in effect a leafy peduncle with a terminal inflorescence. However, the shoot is more commonly quite short and the leaves below the catkin are then small and bract-like; or, as in the sallows and many other species, the peduncle is so much reduced that the catkin is to all intents sessile in the scars of the previous year’s leaves. In some alpine species of the subgenus Chamaetia there are no specialised flowering shoots: the catkin is terminal on a normal growth equipped with both leaves and axillary buds. See also S. exigua for the inflorescence of the Longifoliae. The willows are dioecious, i.e., the catkins are composed of either male or female flowers, and the individual plant bears catkins of one sex only. Rarely and abnormally (especially in hybrids) some flowers are bisexual, or some catkins bear both male and female flowers; see further under S. aegyptiaca and S. × sepulcralis. The catkins vary much in length, shape and relative width. They may be produced before the leaves (precocious) as in the sallows, or in such species as S. daphnoides and S. viminalis; with the unfolding leaves (coaetaneous); or during summer after the leaves are expanded (serotinous), as in S. triandra and S. pentandra.
Each flower is subtended by a scale (sometimes called a bract), which is entire or, very rarely, slightly toothed, usually persistent, long-hairy, sometimes merely downy or shortly ciliate, rarely quite glabrous. The scale may be uniformly coloured, or reddish or pinkish at the tip, or, as commonly in the subgenus Vetrix (Caprisalix), dark brown, dark purple, or almost black in the apical part. Between the flower and the axis of the catkin there is a minute fleshy body usually termed a nectary (but in older works called ‘gland’). In many species there is a second nectary between the flower and the subtending scale, at least in the male flowers, and sometimes the two nectaries are united into an irregularly lobed disk (the nectaries are in fact homologous with the disk seen in the flowers of Populus, and those willows that have united nectaries usually show other ‘poplar characters’). The nectaries are too minute to be of much use for identification, except to the specialists, and are omitted from the descriptions of most species in this work.
The flowers have no sepals or petals (though some botanists consider the nectaries to represent a much reduced perianth). The male flowers have mostly two stamens, but up to twelve in the mainly subtropical section Humboldtianae, and three to five in some other members of the subgenus Salix. In S. purpurea and a few other species (some closely related to it) the filaments are completely connate and the stamen apparently solitary. The female flower consists of a single ovary, which may be sessile or borne on a ‘stalk’ (strictly a pedicel or stipe). The ovary is variable in shape, and may be glabrous, silky or coated with matted hairs. The style is single, or occasionally forked in the upper part, and is very short or lacking in some groups. There are two stigmas, each entire or more often retuse or more or less deeply divided into two lobes or slender arms.
It is an interesting fact that the flowers are insect-pollinated (many willows are indeed useful bee-plants). In this respect salix differs not only from most other catkin-bearing genera (beech, oak, hazel, etc.) but even from the allied poplars, all of which rely on the wind to spread their pollen. An exception is S. arbutifolia (often separated from Salix as the genus Chosenia) and it is said that some dwarf willows of high latitudes are partly wind-pollinated. The fruit, which usually ripens very quickly, is a capsule splitting in two at the top, each half recurving. The seeds are minute, with a tuft of pale hairs at one end. They soon lose their vitality but germinate rapidly if they fall on favourable ground. It is said that seeds of the American black willow (S. nigra) grow 4 ft high by autumn.
The genus is widely distributed in the northern hemisphere. The section Humboldtianae is mainly tropical and subtropical and extends into the southern hemisphere. At the other extreme, dwarf willows extend farther towards the North Pole, and farther towards the eternal snow, than any other woody genus. In this respect the willows have proved themselves more adaptable and adventurous than the allied poplars.
Many natural hybrids have been described in Salix, but they occupy more space in the literature of the genus than they do in the wild, and with few exceptions (e.g., S. × rubens) are never so frequent under natural conditions as to blur the boundaries between species. They rarely occur in closed plant communities, even where, as in the Alpine valleys, four or five species may grow in the same locality, and are most likely to be found in such habitats as road- or railway-embankments, landslides, new scree or waste ground, where the more highly adapted parents may be at a disadvantage. Many hybrids that have been described, or identified in herbarium annotations, are certainly no more than forms of variable species. Either the writer had only a limited knowledge of the species; or, though knowing his subject well, he nevertheless defined the species arbitrarily, calling on hybridity to explain variations that lay outside the subjective boundaries he had laid down.
Some hybrids, though occurring rarely in the wild, or even uniquely, have been brought into gardens and multiplied there. Of greater importance are the hybrids of the various species used in basketry, many of which were widely planted in the osier-grounds and for estate use, thus acquiring so extensive a distribution as to be taken for species by early salicologists such as Sir James Smith, who denied that hybrids ever occurred in the genus. Hybrids have also been raised artificially by researchers and enthusiasts. It is interesting that female plants often far outnumber males in the progeny of these crosses.
For the natural hybrids occurring in Britain see: R. D. Meikle, Salix, in C. A. Stace ed., Hybridization and the Flora of the British Isles (1975), pp. 304–38.
Since prehistoric times the stems of the willows have been put to innumerable uses, from crude hampers and hurdles to the finest sort of basketware. Great osier-beds once stretched along the Thames from Reading downwards to supply the London basket-makers, and today there are still extensive beds in areas so damp or so subject to inundation as to be unavailable for ordinary crops. But some of the finest quality wicker-work is made from willows grown on ordinary farm land. Often several forms of osiers used in basket-making are derived from a single species. These clones, although they vary much in quality for their particular purpose, show no significant botanical differences, and are known in the osier trade by colloquial names. Some of these clones were introduced from the continent within the past few centuries, but some may be very ancient, for the Celts were skilled basket-makers and taught their craft to the Romans.
The value of several sorts of willow trees for producing wood from which cricket-bats are made is alluded to under S. alba ‘Caerulea’, S. fragilis and S. × rubens. But these, as well as S. alba itself, are useful also for other purposes, especially where a non-splintering wood is required, and where it is subject to rough friction as on cart or wheelbarrow bottoms. For wattling the wasting banks of rivers or other pieces of running water nothing equals the branches of willows. Some of the tree species or the taller sallow hybrids are useful for fast-growing shelter-belts. All the willows are good bee-plants, the early-flowering kinds being the most valuable. Altogether the willows are a useful and undemanding race. In his edition of Miller’s Dictionary (1807) Thomas Martyn summed up the long service of the willows to man when he wrote: ‘The antient Britons made boats of wicker covered with skins (coracles) by which they passed rivers and arms of the sea; so light as to be carried by a single man. Modern Britons wield bats of Willow, in the game of Cricket.’
Most willows are propagated extremely easily by means of leafless cuttings, which may be put in the open ground at any time between November and early March. Pieces one to several years old may be used; and of the tree sorts like S. alba, S. fragilis and S. × rubens, it is usual to put in ‘sets’, i.e., naked rods, 8 to 12 ft long, and as thick as, or thicker, than a broomstick. But cuttings of half-ripened wood, taken from June onwards, will also root readily and this is the most convenient method of increasing the dwarfer species. Only a few willows are hard to strike, one notable example being male plants of pure S. caprea. Plants raised from cuttings in the nursery should be put in their permanent places at not more than two years old and should be planted a few inches deeper than before; the best results will, indeed, be obtained by putting the cutting into its destined place at the start if due protection and care can be given. This is always done with the big ‘sets’ just mentioned.
The majority of willows abhor dryness at the root, but will thrive in ordinary situations if the soil be deep. The sallows will grow well on poor, unimproved soils and make useful pioneers and soil-fixers. S. purpurea too is tolerant of dry soils, while S. daphnoides has been used on the continent for the fixing of sand-dunes. A few species and hybrids of the Alpine type are suitable for the rock garden and some of the more vigorous ones make useful ground-cover.
The willows are subject to several fungal diseases with similar symptoms, all causing die-back of the young growths and lesions on the older stems. These have been studied mainly in relation to the clones grown in osier grounds, since the cankering and discoloration they cause render the rods unfit for basketry. In gardens the most serious damage is likely to be suffered by the various hybrids of S. babylonica, whose pendulous growths may be greatly shortened by a severe attack. Both Willow Anthracnose and Willow Scab can cause this damage, and the former is most prevalent in a wet spring.
The bacterial Watermark disease has caused severe damage in plantations of the Cricket-bat willow, S. alba ‘Caerulea’, but is mainly confined to areas where this is grown extensively. Even if not lethal, an attack causes staining in the wood of the trunk and makes it commercially valueless.
The attacks of a Gall Mite cause the catkins of some willows, especially the sallows and their allies, to become woody and persist on the branches as unsightly protuberances. The form or hybrid of S. purpurea once known as S. helix bore the vernacular name ‘rose-willow’ from the curious rosettes of short leaves sometimes formed at the tips of the branchlets, and the same abnormality has been observed in other willows. It is caused by the punctures of a species of Gall Midge.
It is hoped that the following synopsis may help to make this large and confusing genus a little more comprehensible. It is largely based on A. K. Skvortsov, Ivi SSSR (Willows of the USSR) (1968), a taxonomic and geographic survey, with keys, that covers all the Old World willows with the exception of some little studied groups largely confined to China and the Himalaya. Most of the American species have been fitted into this classification with the aid of some recent works, notably: Robert E. Dorn, ‘A synopsis of American Salix’, Canadian Journal of Botany, Vol. 54 (1976), pp. 2769–89; George W. Argus, The Genus Salix in Alaska and the Yukon (1973); Arthur Cronquist, Salix, in C. L. Hitchcock et al., Vascular Plants of the Pacific Northwest, Part 2 (1964).
Trees or large shrubs. Leaves elliptic or lanceolate, much longer than wide, acute or acuminate, toothed. Catkins slender, sometimes drooping; scales uniformly coloured, usually deciduous in the fruiting catkins. Nectaries in male flowers two (sometimes more numerous and then united into a disk); in female flowers one or two. Stamens free, two or mote.
sect. Humboldtianae (Nigrae) – Buds triangular, not appressed to the shoot, their scales free from each other on the inner (adaxial) side. Nectaries in male flowers sometimes more than two and then more or less connate at the base. Stamens three to ten, with small anthers. Ovary stalked; stigmas very short, sessile or on a short style. A group of about a dozen species, mostly in the tropics and subtropics, but in N. America extends in the cold temperate zone; see S. nigra. In western Eurasia its northernmost representative is S. acmophylla Boiss., a native mainly of Iran and bordering parts, but extending as far west as Palestine and S.E. Turkey. S. humboldtiana Willd., of Central and S. America, would probably be hardy in Britain; it has a fastigiate form cultivated in the warmer parts of America and also in New Zealand.
sect. Amygdalinae – A group of two or three species in the Old World. See S. triandra.
sect. Urbanianae – The one (perhaps two or three) species in this E. Astatic section are mentioned under S. arbutifolia.
sect. Pentandrae – Trees or large shrubs. Leaves glossy, glandular-serrate, often balsam-scented when young, usually with well-developed glands at the junction with the petiole. Catkins borne after the leaves unfold, on leafy twigs. Catkin-scales usually glandular at the apex. Nectaries two (sometimes more numerous in the male flowers and then connate into a disk enclosing the bases of the stamens). Stamens three to ten. Ovary stalked; style very short; stigmas bilobed. About eight species in Eurasia and N. America. Those treated here are S. pentandra and the American S. lucida and S. lasiandra. Apart from S. caudata, which is doubtfully distinct from S. lasiandra, the only other American species is S. serissima Fern., a close ally of S. pentandra remarkable for producing its catkins in autumn.
The E. Asiatic section Glandulosae, probably not represented in cultivation in Britain, is intermediate between the Humboldtianae and Pentandrae. The type is S. chaenomeloides Kimura (S. glandulosa Seem., not Raf.) a native of Japan, Korea and China.
sect. Salix (Amerina, Albae) – A group of three species in W. Eurasia. See S. alba (here taken to be the type-species of Salix) and S. fragilis. The third species is S. excelsa Gmel., a native of S.W. Asia. In this section the stamens are reduced to two
sect. Longifoliae – There are about five species in this very distinct section, natives of N. America and Mexico. See S. exigua.
sect. Subalbae – A group of a few species in E. Asia, often regarded as the counterpart of the more western S. alba and S. fragilis. They differ from most other species of the subgenus Salix in the very short catkins with persistent bud-scales and in having the catkin-buds of different form from the growth-buds. See further under S. babylonica, S. matsudana and S. jessoensis.
Of creeping or procumbent habit or erect but dwarf. Leaves always relatively broad. Catkins on leafy laterals, the leaves on these being of the same size as those on the vegetative growths, and as numerous. Nectaries large, often more than one, sometimes lobed. Catkin-scales persistent. Stamens two or three, rarely solitary. All the sections are represented in the arctic and boreal regions of Eurasia and America, and many of the species occur at high altitudes in mountains farther south.
sect. Chamaetia (Reticulatae) – See S. reticulata, the only European representative, to which the American S. nivalis Hook. (S. saximontana Rydb.) is closely allied. Of the other two species S. vestita Pursh is American but also occurs in the Altai Mountains of Russia.
sect. Retusae (Herbaceae) – Cushion-forming or prostrate shrubs, sometimes with underground stems. Stipules inconspicuous or wanting. Catkins terminating normal leafy shoots. Filaments of stamens glabrous. Ovary narrow, tapered into the style; stigmas small. See S. herbacea and S. retusa. The Japanese S. pauciflora Koidz. belongs to this section.
sect. Myrtosalix – Small erect shrubs, more rarely prostrate. Leaves glossy above, often persisting withered on the branches. Stipules usually present, symmetrical. Catkins terminating leafy shoots, their scales dark-coloured. Ovaries glabrous or clad with matted hairs; style well developed. See S. myrsinites and S. uva-ursi. The latter is an American species. So too is S. dodgeana Rydb., a rare native of the Rocky Mountains, closely allied to S. rotundifolia Trautv. of the Aleutians, Alaska, etc.; it is a mat-forming species with minute leaves, often found in hollows where the snow lies long.
sect. Glaucae (Ovalifoliae) – Prostrate or small, erect shrubs, rarely medium-sized. Leaves entire or faintly toothed; stipules narrow or wanting. Catkins on leafy stalks, but these often markedly shorter than the vegetative shoots. Nectaries sometimes more than two in the male flowers, the adaxial nectary often large and lobed. Ovaries stalked, with a well-developed style. See S. arctica, S. glaucosericea and S. pyrenaica. The small section Myrtilloides, see S. myrtilloides, is related to the Glaucae.
Also belonging to this subgenus, but scarcely represented in British gardens, is the section Lindleyanae, for which see S. lindleyana.
subgen VETRIX (Caprisalix)
To this taxonomically difficult group belong about two-thirds of the species of Salix. They are shrubs or small trees. Catkins usually emerging before or with the leaves; scales dark, at least at the tip. Nectaries normally one in both sexes. Stamens two, the filaments sometimes wholly or partly united.
This group is exceedingly variable in foliage. In the great majority of the species the catkin-buds are markedly larger than the vegetative ones, and are produced near, but not extending to, the apex of the shoot, the uppermost buds, as well as those at the base of the shoot, being vegetative.
sect. Eriostachyae (incl. Magnificae) – A taxonomically rather isolated and primitive section, anomalous in this subgenus in having two nectaries in the male flowers. The species are shrubs with large or very large leaves. Catkins on leafly stalks, of a remarkable length in some species. Stamens two. Ovary sessile or shortly stalked, with a short style. The species, number uncertain, are natives of China and the E. Himalaya. See S. fargesii, S. magnifica and S. moupinensis.
sect. Cordatae (Hastatae) – Mostly small or medium-sized shrubs. Leaves finely toothed or entire. Stipules well-developed, symmetrical or almost so, often a conspicuous feature. Catkins usually with or shortly after the leaves. Ovary glabrous, acute, with a definite style; stigmas small. A northern and alpine group, mainly American. See S. cordata, S. hastata, S. rigida. The American section Balsamiferae is included in the Cordatae by some authorities; see S. pyrifolia, the only species in this section.
The new section Glabrella, set up in 1968, comprises some species previously grouped with S. hastata; see S. glabra.
sect. Nigricantes – A very small group, confined to the Old World. See S. nigricans.
sect. Vetrix (Capreae) – Shrubs or sometimes small trees; wood under the bark often striated (with raised ridges). Stipules well-developed in most species, asymmetrical. Leaves unequally toothed or entire, broad. Catkins almost always produced before the leaves. Stamens two. Ovary stalked, the stalk often elongating as the capsule matures; style very short. A large group in the temperate regions of the northern hemisphere, mostly in the Old World. See S. aegyptiaca, S. aurita, S. caprea, S. cinerea, S. discolor, S. silesiaca, S. starkeana.
sect. Arbuscella (Phylicifoliae) – Shrubs, sometimes dwarf. Catkin-buds not markedly different from the vegetative. Leaves discolorous, being dark green above, white or grey and finely reticulate beneath. Catkins before, with or after the leaves. Ovary clad with appressed hairs, tapered into a definite style. A group of about eighteen species in the northern parts and mountains of the Old World, a few in N. America. See S. arbuscula and S. phylicifolia.
sect. Vimen (Viminales) – Trees or large shrubs. Stipules present, of various shape. Petioles subtending the catkin-buds much expanded at the base. Leaves narrow, with numerous parallel lateral veins, entire or finely toothed, silky beneath at least when young. Catkins usually precocious. Stamens two. Ovary short-stalked or sessile, with a definite style; stigmas divided. A group of about ten species in the Old World. See S. sachalinensis and S. viminalis; also S. dasyclados, p. 306.
sect. Subviminales (Gracilistylae) – Two or three species, closely allied to the section Vimen, but with the filaments of the stamens partly or wholly united. See S. gracilistyla. Another close ally of the section Vimen is S. elaeagnos (q.v.), the only species of the section Canae; in this species too the stamens are partly united.
sect. Villosae – Shrubs (one species with a short trunk). Leaves entire or obscurely toothed, their undersides and the stems white-woolly. Catkins usually before the leaves, thick and very hairy. Ovary sessile or short-stalked; style elongate. About five species in the colder parts of the N. hemisphere. See S. candida and S. lapponum. This section has some affinity with Vimen, indeed S. lapponum is grouped with S. viminalis by some authorities.
sect. Lanatae (Chrysanthae) – This group of about five species bears much resemblance to the preceding, but the undersides of the leaves have a greyer, less matted indumentum and the stipules are generally larger and almost symmetrical. In the American species the uppermost buds of the shoot are usually all catkin-bearing, the uppermost catkin apparently terminating the shoot. See S. lanata. The N.W. American S. richardsonii has recently been reduced to the rank of a subspecies of S. lanata by A. K. Skvortsov. S. barrattiana Hook. ranges from Alaska to British Columbia, while the rare S. tweedyi (Bebb) Ball (S. barratiana var. tweedyi Bebb) has a more southern distribution.
sect. Daphnella (Pruinosae) – See S. daphnoides, under which are mentioned the other three species in this section.
sect. Incubaceae – A group of probably only three species, these confined to the Old World. See S. repens.
sect. Helix (Purpureae) – Shrubs or small trees, with slender, flexible branches. Leaves usually narrow, not revolute, venation not prominent. Catkins slender, mostly before the leaves in the cultivated species. Stamens two, their filaments partly or wholly connate. Styles short or wanting, with small stigmas. A group of nearly thirty species confined to the Old World, the majority in Central and E. Asia. See S. caesia and S. purpurea.
An important addition to the list of works cited on page 250 is: R. D. Meikle, Willows and Poplars of Great Britain and Ireland, published in 1984 by The Botanical Society of the British Isles (BSBI Handbook No. 4), admirably illustrated with botanical drawings by Victoria Gordon. This provides a succinct and authoritative account of the native species of Salix and their natural hybrids.
An account of the willows occurring in British Columbia, many of wider distribution in North America, will be found in the work by T. C. Brayshaw cited in this supplement under Betula.