kinds which most resemble each other, not only may the species having a close natural affinity be recognised at a glance, but even the tyro will be greatly assisted in tracing and identifying his specimens. If, however, the usual arrangement of the species be adopted, in which the sections are characterised by having the ovaries naked or pubescent; the leaves glabrous or downy, serrated or entire [as in Smith's English Flora, and the Sal. Wob.]; then species widely separated by nature and habit must necessarily be grouped together, not to mention that these characters are in themselves liable to great changes. Fries (in Syllog. Nov. Pl. Soc. Bot. Nat.; Ratisb. edita, t. 2. p. 36.) first distributed the Swedish species of this genus into natural groups, according to characters taken from various parts of the plant. In like manner, I have attempted a similar distribution of the European species; but, first, I shall offer a few words with respect to the characters according to which I have divided the genus into sections and species.

"A character taken from the catkins appearing earlier than, at the same time with, or later than, the leaves is of great importance; but one taken from the situation and insertion of the catkins is still more so. The situation may be in three different modes. 1. In this a catkin is produced at the tip of a branchlet, with a few others below it, and they are all sessile; the leaves proceeding from buds at the base of the catkins. I only know of one instance of this, S. lanàta. 2. A bud on the tip of the last year's branchlet puts forth a catkin, and the peduncle on which it is situated increases in size, and bears leaves, in the axils of which are the buds of the following year. This peduncle is, therefore, persistent, and continues the branch. This is the case in S. reticulata, S. herbàcea, S. polàris, S. retùsa, and S. U`va-úrsi. 3. A terminal bud, and generally more protruded beneath it, produce leaf-bearing shoots, and the flower buds are situated beneath these. All the other species which are known to me, except those enumerated above, belong to this division; and they may be subdivided as follows:-1. Those in which the catkin is sessile, on a very short peduncle, or as it were incipient, and bears at its base weak scale-like leaves; being thus lateral, sessile, and bracteated at the base. 2. Those in which the peduncle grows into a branchlet, and bears floral leaves not very distant from the catkin, which afterwards become true leaves, but without buds in their axils: from this branchlet is formed the lateral catkin, which is peduncled with a leafy peduncle. All the species which protrude their catkins before their leaves belong to the first of these subdivisions; and all those which do not protrude their catkins till after their leaves, with many of those which protrude their catkins at the same time as their leaves, to the second. This character seldom changes; and only a few species (for example, S. limosa) bear on one plant, or, as a variety, on two plants, catkins which have short peduncles, and are surrounded at their base with very minute scale-like leaves; and also those that are peduncled, and have true leaves on their peduncles. Even in these varying forms Nature shows her inexhaustible fertility, and her wonderful skill and power of adaptation in creation: despising the too great carefulness of learned men, who hasten to build prisons for their own systems, she delights in disturbing their magic circles, and, playfully breaking loose from the chains in which they have attempted to bind her, she far exceeds Proteus himself in versatility.

"The importance of the characters which the pedicel of the capsule offers has been pointed out by Wahlenberg. Its length relatively to the gland, which is never wanting, is a very constant character, varying only in a few species; but, to be rightly observed, it ought to be seen just at the time when the ovary attains the size of a capsule, which happens a little after flowering; or, in dried specimens, if accuracy is wanted, part of the female catkin must be softened in boiling water, and afterwards dried in blotting-paper, before examination. In dried specimens, the pedicel is so brittle, that in the analysis it is seldom preserved entire; or, from being joined to a gland not less fragile, it is frequently injured. Besides, it must be remarked, that some catkins have been found in which the inferior flowers were very remotely situated.

In these instances, the pedicel is often a little longer, and the capsules more slender.

"The colour of the young shoots varies greatly, often so much as to cause the varieties to appear distinct species. The branchlets of S. álba are either brown, or, as in the var. vitellina, of a yolk-of-egg or a red brick colour; and there is a different shade of yolk-of-egg colour in S. rèpens, and S. rosmarinifòlia (S. læ`ta Schultz). Many species, when carefully examined, will be found to vary in colour, though only to a small extent. The branches of S. purpurea are of a coral colour, rarely of a dark yellow, and sometimes white, covered with a reddish bloom. S. amygdálina has the shoots sometimes of a brownish yellow, and sometimes of a brownish black.

"The form of the leaves in the same species, and even in the same plant, can never be depended upon. In S. phylicifòlia, S. myrtillöìdes, S. arbúscula, and S. rèpens, they vary from narrow-lanceolate and being attenuated towards the base, in the three last-named species, to roundish-ovate and being cordateemarginate at the base. In some species, the form of the leaves is almost always the same, as in S. viminalis, S. incàna, and S. hippophaefòlia. In other species, the leaves vary; being serrated or entire, green or hoary on the under surface, and glabrous or hairy, on the same plant. The same variation is common on the exterior of the ovaries; which, in S. phylicifòlia, are sometimes glabrous, and sometimes hairy; some individuals of this species having half the ovary hairy, and the other half glabrous; while in others there is only a hairy or downy line. In certain species, however, these variations are never found, or very rarely; although in S. viminalis ovaries partly naked, and partly downy, occur. The brown tip of the bracteas of the flowers, in some species, turns paler, and in others red, or even purple; which is another cause of uncertainty in specific distinctions. The bracteas are sometimes obovate, and only half the length of the ovary; and sometimes, in the same species, lanceolate, and reaching as far as the style. The style and stigma likewise vary in length, and are occasionally more or less cleft; yet both these organs afford most useful characteristics. The style often appears shorter from being hidden by the long hairs of the ovary. Stigmas of a rose colour, and of a yellow colour, have been found in the same species. The stipules vary in size, but never in form; hence they afford the very best characteristics for distinguishing species. In no species can these be said to be wanting; and, though on old plants they are often not seen, such plants, when cut down, send up young shoots which produce leaves attended by stipules of an extraordinary size. The buds are always l-valved; and the valves are often cleft at the tip, and sometimes as far as the base; though sometimes, on the same individual, they are undivided. The folding of the leaves in the bud is, most probably, constant, although different in the various species: but this I cannot affirm as certain, not having examined the leaf buds of a sufficient number of species. "The variation of the different parts is not the only difficulty with which the botanical student, in this genus, has to contend: the great number of hybrids, the existence of which in the genus Salix no one can doubt, is another obstacle. Nobody will accuse me of arrogance in assuming to know S. rùbra and S. viminalis. On the banks of the Rednitz, near Erlangen, there are many thousand trees of these two species; and, at the same time, many intermediate forms, which I can refer to neither species. The catkins of these afford no distinguishing marks; for what seem at one time to belong to the former species, at another time appear more nearly allied to the latter." Koch concludes by stating that, in his Commentary, the species have been arranged in 10 groups; and that no kind has been admitted as a species that he has not himself seen and examined. He has added but few varieties," although an immense number of no importance might have been adduced; being convinced, from daily observation and experience, that the multiplication of varieties, instead of rendering any intricate genus more clear, only involves it in a greater difficulty."

The species of Koch, besides being identified with those of the Species

Plantarum of Linnæus, and the Species Plantarum of Willdenow, have the synonymes of other authors added to them.

In our App. iii. to the genus Sàlix will be found the characters of Koch's 10 different groups; and under each the names and synonymes of the species which he has assigned to them.

From the perusal of Koch's observations, two points, we think, will be rendered clear to the botanical reader: -1. That the mode of arranging the sections according to the character of the leaves, adopted by all the Linnæan school previously to the time of Wahlenberg, is altogether defective; and, 2. That the system of throwing the species into natural groups, as adopted by Wahlenberg, Fries, Koch, and Borrer, is the true one. Being ourselves of this opinion, the only question that remained for us to decide was, whether we should follow Koch or Borrer in the arrangement of the species described in this work as in a living state in British gardens.

The excellence of Koch's system was strongly impressed on our mind from the moment that we saw it developed in Dr. Lindley's Synopsis of the British Flora; and, if we could have classed all the numerous sorts of willows m the salictum at Woburn, and in the Hackney arboretum, under Koch's ten groups, in a manner satisfactory to ourselves, we should have done so; the more especially as, from observing with care all the different sorts in the Hackney arboretum, at different periods, from March to December, 1836, we felt convinced in our own mind that by far the greater number of them were varieties, and chiefly of S. càprea L. Not being able to do this, we determined on endeavouring to obtain the advice and assistance of the first authority in Britain on the subject of willows; and we accordingly applied to Mr. Borrer, who at once, in the most kind and liberal manner, classed the sorts contained in the Salictum Woburnense in the 22 groups into which, with the exception of a few sorts, they are thrown in the following article. Mr. Borrer's knowledge of this genus is universally known. He possesses an extensive collection of living plants, which he has cultivated for some years; and, as Sir W. J. Hooker remarks," No one has ever studied the willows, whether in a growing or a dried state, more deeply, or with a less prejudiced mind." (Brit. Fl., ed. 3., vol. i. p. 416.)

The botanical details which we have given of each particular species, including a comparison of specimens obtained in a living state from the arboretum at Flitwick, from that at Goldworth, and from the salictum at Messrs. Loddiges's, were made out for us, with great care and industry, by Mr. Denson. Our figures were chiefly drawn for us by Mr. Sowerby, from specimens received from the salictum at Woburn Abbey; in the single instance of the S. càprea, reduced from Host's work; and nearly all the remainder, including all the 28 plates of leaves of the natural size, by the kind permission of the Duke of Bedford, have been copied from the Salictum Woburnense.

It will thus appear that our article, lengthy and elaborate as it is, is, in a botanical point of view, chiefly to be considered as matter for a history of willows, rather than as a complete history in itself. Such a history, indeed, can only be prepared by a botanist who has had all the species in a living state under his eye for several years; and who has applied to them one general principle of contrast or comparison. Till this is done, not only with the genus Salix, but with every other genus of which there are numerous species, a decided imperfection must ever be found in works like the present, in which the specific characters are necessarily made up of descriptions given by different individuals, at different times, and in different countries; some from living plants collected from their native habitats, others from living plants grown in gardens, and many from dried specimens. All this shows the great advantage that would result to botany and arboriculture from a national arboretum; in which not only all the species and varieties should be collected, but also both the sexes of all the kinds that have the male and female flowers on different plants. Such an arboretum, on a sufficiently large scale, and properly managed, would form a living standard of reference, both for the botanist and the cultivator.

Group i. Purpurea Koch, Borrer.
Osier Willows, with one Stamen in a Flower.

Monandra is the name adopted for this group in Hook. Br. Fl., ed. 3.; but Mr. Borrer considers Purpurea preferable, because it is taken, like the name of each of the other groups in this arrangement, from the name of a species included in that group. Purpurea, too, is the name given by Koch to the same group.

Filament 1, bearing an anther of 4 lobes, and 4 cells; or, in S. rùbra, forked, and each branch bearing an anther of 2 lobes and 2 cells. Germen sessile. Catkins very compact. Trees of low stature, or shrubs with twiggy branches, and leaves that are more or less lanceolate, and serrated, and often broader upwards. Interior part of the bark, in most, yellow and very bitter. (Hook. Br. Fl.) The leaves of nearly all of the kinds of this group turn black in drying. The inner bark of most of the kinds included in this group is extremely bitter, which renders the plants suitable for banks of rivers, and other places which are infested by rats; as the bitterness prevents these animals from eating it.

1. S. PURPUREA L. The purple Willow.

Identification. Lin. Sp. Pl., 1444.; Smith Eng. Bot., t. 1388.; Eng. Fl., 4. p. 187.; Forbes in Sal. Wob., No. 1.; Hook. Br. Fl., ed. 3., p. 417.; Mackay Fl. Hibern., pt. 1. p. 243.; Hayne Abbild., p. 229.

The Sexes. Both sexes are figured in Eng. Bot., and are in cultivation in some English collections. Synonyme. S. purpurea a Koch Comm., p. 25.

Engravings. Eng. Bot., t. 1318.; Sal. Wob., No. 1.; Hayne Abbild., t. 169.; our fig. 1294.; and fig. 1. in p. 1603.

1294

Spec. Char., &c. Branches trailing, decumbent. Leaves partly opposite, obovate-lanceolate, serrated, very smooth, narrow at the base. Stamen 1. Stigmas very short, ovate, nearly sessile. (Smith Eng. Fl.) A native of Britain (between Thorpe and Norwich, &c.); flowering in March and April. In a wild state, this species forms a shrub, with a stem 3 ft. or 4 ft. high, with long, slender, smooth branches, spreading widely, and, if not supported, trailing on the ground; very smooth, of a rich and shining purple, with a somewhat glaucous hue. The catkins appear earlier than the foliage; and often on different branches. In cultivation, in dug grounds kept moist and the plants cut down yearly, this species produces shoots from 3 ft. to 5 ft. long, which are much esteemed for the finer sorts of basketwork. It is also frequently planted in Norfolk and Suffolk, and in some parts of Essex, for "plaiting into close low fences, for the exclusion of hares and rabbits; the bark and leaves being so extremely bitter, that these animals will touch neither; whilst the shoots, being long, tough, and flexible, may be formed into any shape; and a fence of this kind is reckoned little inferior to that of wire." (Eng. Flora, quoted in Sal. Wob., p. 2.) This species is well adapted for planting in ornamental shrubberies, from the elegant slenderness of its twigs during winter; the redness of its catkins, the anthers of which are of that colour before they burst, and the fine purplish and glaucous hue of its young shoots and leaves. The latter, as will be seen by the figure of one of the natural size in p. 1603., are of an elegant, and, if we may use the expression, artistical shape. Female plants are in the Hackney and Goldworth arboretums, and at Woburn and Flitwick; and male and female at Henfield. The male plant, being the most beautiful when in flower, ought to be most propagated by nurserymen. Varieties. Koch, in his De Salicibus Europais Commentatio, has described six; but he includes the S. Hèlix and Lambertiana (to be described as species below) as two of them. He has characterised the six varieties as follows:

S. p. 1; S. purpurea Smith, Willd. Stem dwarfer. Branches more spreading. Catkins very slender.

S. p. 2; S. Lambertiana Smith, Willd. - Catkins twice as stout, and leaves larger and broader than in S. purpurea; otherwise not different.

S. p. 3; S. Hèlix Willd. En. -Branches uprightish, but spreading. Leaves longer.

S. p. 4 monadelphica. A male plant, with the stamens divided to the middle, or, rather, having 2 stamens with the filaments connate, as in S. rubra, and as far as to the middle. Koch found this growing in the Palatinate of the Rhine, near Cassel.

S. p. 5 sericea; S. monándra sericea Ser. Sal, Helv., p. 8.-This has its leaves, while they are young, covered with a dense silky down, which afterwards disappears. Seringe observed this in Switzerland; and Koch afterwards gathered it in the Palatinate.

S. p. 6 bráctea rùbra.-This has the scales of the catkin, that is the bracteas, of the colour of red brick, and not black. Günther sent it to Koch from Silesia; and Koch deems it a rare and singular variety.

Remark. Koch, considering S. purpurea as including the above four, gives the geographical dis tribution of the species as follows:- It inhabits the banks of streams and moist meadows, and also sandy and comparatively dry places, in plains and lower mountains, from the Pyrenees and Alps, through England and the whole of Europe, as far as to the south of Sweden.

Identification. Lin. Sp. Pl., 1444.; Willd. Sp. Pl., 4. p. 672.; Hayne Abbild., p. 229. t. 170.; Smith Eng. Bot., t. 1343.; Eng. Fl., 4. p. 188.; Forbes in Sal. Wob., No. 2.; Hook. Br. Fl., ed. 3., p. 417.; Mackay Fl. Hibern., pt. 1. p. 244.

Synonymes. S. purpurea var. Koch. Comm., p. 25.; ?S. oppositifòlia Host Sal. Austr., 1. p. 11. t. 38, 39.

The Sezes. Both sexes are figured in Sal. Wob., and also in Eng. Bot.; but Mr. Borrer believes that the catkins of female flowers represented in the latter are those of S. Forbyàna: if those of Helix, they are much too thick. Mr. Borrer having only seen the male of S. Hèlix, and the female of S. Lambertiàna, is inclined to regard them as the two sexes of one species.

Engravings. Eng. Bot., t. 1343., the male plant; Sal. Wob., No. 2.; Hayne Abbild., t. 170. ; and fig. 2. in p. 1603.

Spec. Char., &c. Branches erect. Leaves partly opposite, oblong-lanceolate, pointed, slightly serrated, very smooth; linear towards the base. Stamen

1. Style nearly as long as the linear divided stigmas. (Sal. Wob., p. 3.) A native of Britain; flowering in March and April. A tree of humble growth, but erect; about 10 ft. high, smooth in every part, altogether of a lighter hue than those of S. purpurea. The branches are not trailing, but upright; they are smooth and polished, of a pale yellowish or purplish ash colour, tough and pliable; less slender and elongated than the foregoing, though useful for the coarser sorts of basketwork. Catkins larger than those of S. purpurea; the fertile ones, especially, full twice as thick. (Eng. Flora, p. 188.) The branches, which are yellow, and the mode of growth, which is erect, render this species easily distinguishable from the preceding. Description. The name rose-willow relates to rose-like expansions at the ends of the branches, which are caused by the deposition of the egg of a cynips in the summits of the twigs, in consequence of which they shoot out into numerous leaves, totally different in shape from the other leaves of the tree, and arranged not much unlike those composing the flower of a rose, adhering to the stem even after the others fall off. (Smith, and Kirby and Spence.) Smith had never seen this monstrosity but on S. Helix, except once on S. aurìta: but it is very common on S. Hoffmanniana in Sussex (Borrer), and on S. álba in Cambridgeshire, and is obvious in winter when the plants are leafless. In these two kinds, the rose-like bodies are constituted of leaves imbricately disposed, the upper the smaller: some of the bodies are 3 in. over. "The leaves and twigs are less bitter than those of S. purpurea; and the greater size of the stem, as well as branches, renders this species fit for several purposes which that is not. It also makes a better figure in plantations, and the roots give more solidity to the banks of rivers or ditches." (Smith.) Gerard describes the rose-willow, of which he has given a figure, as "not only making a gallant show, but also yielding a most cooling aire in the heat of summer, being set up in houses for the decking of the same." Dr. Johnston, in his Flora of Berwick upon Tweed, states that S. Hèlix withstands storms better than any other species. A crystallisable principle, called salicine, has been obtained from this species; which, according to Majendie, arrests the progress of a fever with the same power as sulphate of quinine. (Jour. R. Inst., October, 1830, p. 177.; Lindl. Nat. Syst., p. 187. See also our p. 1459.) In ornamental plantations, S. Hèlix is an interesting shrub, from its