arrangement of vegetables, founded on that of Cæsalpinus and Morison, consisting of twenty-five classes, which he afterwards. improved and republished in 1700. His improved method consisted of thirty-three classes, which is divided into two grand divisions, viz. plants destitute of buds, i. e. herbs; and plants producing buds, i. e. trees. Herman, professor of botany at Leyden, Christopher Knaut, the celebrated Boerhaave, and many others, formed systems, by altering and attempting to improve those of their predecessors.

When any set of objects is too numerous to be comprehended by the mind at once, they may, by means of marks of distinction or resemblance, be divided or united into smaller assemblages, which (if the analytic method be followed) may be again divided into subdivisions, or orders; these again into genera, and genera into species or, if the synthetic method be adopted, species may be assembled into genera, &c.

There are two ways of accomplishing this; the one, by uniting into the same assemblage, such species as, from an agreement in several particulars, seem to be connected by a close affinity established by nature; the other, by forming into arbitrary associations a number of species which happen to agree in some accidental circumstances. The former is called natural, the latter artificial method. All the authors of systems we have hitherto mentioned, adopted the natural method. In the vegetable kingdom, it frequently happens, that a great number of species agree in so many particulars, as evidently to demonstrate them to be members of the same natural family. Among the grasses, for instance, the family resemblance, the similarity in the form of the leaves, in the construction of the stalks, in the parts which compose the flower, and in the nature of the seeds, is so great in all of them, as easily to enable any one who may be acquainted with only one or two species, to recognize any other almost at first. sight. There are other tribes of vegetables, the different species of which may be recognized with equal facility, even by those who have but a slight acquaintance with them; but this is by no means the case with all. Though nature generally exhibits a wonderful degreee of regularity, she frequently makes considerable deviations; and anomalous productions are to be found, which, though they possess the principal features of one family, yet, in some particulars, so far resemble another very distinct tribe, as to make it doubtful to which of the two they ought to be referred, without a minute investigation by one well acquainted with the distinguishing characters of both. Nearly allied families, likewise, frequently run into one another so imperceptibly, as to render it no easy task to draw the line of separation.

For

Plantarum flores funt,

For these reasons, it is frequently difficult to define natural families, in such a manner as to comprehend all their irregular members, without at the same time leaving room for the admission of species which do not belong to them.

Whatever may be said, therefore, in favour of natural classes, and however fit they may be, for the use of those who are well acquainted with botany, at a time when the greater part of the ve getables this globe produces, shall have been discovered and described; while they are deficient in point of facility, they cannot be the most proper for beginners, nor could they be so even for botanists themselves, at a period when, comparatively, few plants were known.

A. Quirinus Rivinus, professor of botany at Leipsic, perceiving the defects of the natural systems proposed by his predecessors, endeavoured to form an artificial one, founded on the regularity and irregularity of the corolla, and on the number of the petals of which it was composed, which he published in 1690. It consisted of the eighteen following classes.

This author was the first who ventured to disregard the stamp of antiquity, and rejected entirely the ancient division, which had so long fettered all former framers of systems.

His design was good; but he was unfortunate in the choice of the part of the flower he fixed on for the foundation of his classes, as flowers are more liable to vary in the number of their petals, than in any other particular: he was however more successful in the selection of the fruit for the foundation of his

orders,

orders, because it affords marks, which are subject to little variation.

Tournefort, whose system was more followed than any other, except that of Linnæus, particularly in France, and even in this country, till after the middle of last century, fixed on the form of the corolla for the foundation of his classes, many of which were very natural: but the great variety of forms the corolla assumes, many of which approach each other so nearly, as scarcely to be distinguished, render the application of his system to practice difficult.

Many systems, both artificial and natural, have been devised by different botanists, all of which gave way to the artificial system of Linnæus, termed sexual, because the distinguishing marks of all his classes are taken from the parts of generation, i. e. the pistillum and stamina.

The Linnæan method owes its success, not only to the supe rior facility, as well as certainty, by which any plant, by means of it, may be reduced to its class and order; but to the attention the author has paid to the more minute divisions, and to the pains he has taken to give accurate descriptions of the species, and to affix to each its trivial name. The different publications of Linnæus, particularly his System of Nature, and Species Plantarum, have done more for the advancement of botanical knowledge, than the writings of any other man. Since the time his last

supplement was published, a very great number of species have been discovered in different quarters of the world. Those which have been described, are dispersed through the writings of various authors; others are only to be met with in the herbaria of the curious. Linnæus junior, in his supplement, has described several. Murray and Gmelin, in their editions of the System of Nature, have published only such as they themselves had an opportunity of examining, together with those described by Thunberg and Jacquin.

It was long the earnest wish of those who were occupied with such studies, that some person, properly qualified, would give them in a collected form to the world, Willdenow undertook the Herculean labour; and from the manner in which he has conducted this edition of the Species Plantarum, has shewn himself well qualified for the task. He has adopted none of the alterations in the classes, proposed by the VOL. X. NO. 20. authors

X

* Adanfon, a Frenchman, to fhow the facility with which they may be formed, and his own dexterity, has manufactured upwards of fixty of them.

authors we shall mention below; but has, with two exceptions, adhered very closely to the arrangement of Linnæus, at least in the larger divisions."

Notwithstanding the merits of the Linnæan system, it has without doubt many imperfections, which different botanists have endeavoured to correct, but hitherto without any very distinguished success. The two principal objections are, 1st, that plants, evidently connected by nature, are, according to it, placed in different classes; and that those which have no natural or general affinity, are sometimes arranged under the same class,— an objection to which every artificial system must be liable. 2dly, Plants, by means of it, cannot always be ascertained with such facility and precision as might have been expected in an artificial arrangement. The improvements which have been aimed at, have been suggested principally with a view to make the arrangement more natural; and indeed, Linnæus's own anxiety to remedy its defects in this particular, is one cause of the difficulty which is frequently experienced in reducing a plant to its proper place. In the class Diadelphia, for example, which ought to contain those plants whose stamina are united, by means of their filaments, into two groups, there is a subdivision in the order Decandria, containing a number of genera, characterized Omnia stamina connexa; though it is obvious, that any one who knows merely the general principle of the Linnæan arrangement, would naturally look for such a flower in the class Monadelphia.

Had Linnæus, instead of making this exception, improved the definition of his class by means of the form of the flowers, he would not only have obviated this difficulty, but rendered the class itself more natural, and made it comprehend those irregular members of the family which have their stamina unconnected. For there are many plants standing in the order Monogynia, of the class Decandria, which, from the form of their flower, and nature of their seed-pod, evidently belong to the class Diadelphia.

Any alterations, therefore, made with a view to render the arrangement more natural, which have a tendency to make it more complex, are certainly improper. Instead of depriving an artificial system of its principal recommendation, by patching or twisting it, it would be better, at once, to have recourse to natural classes, such as have been formed by Linnæus himself, by Batsch or Jussieu. But if an artificial arrangement be thought best adapted to the present state of botanical knowledge, such alterations only should be made, as tend to remove ambiguity and difficulty. Of all the alterations and attempts to im

prove the Linnæan system, there are but few that merit notice. Liljebad has made the following changes. He joined the 7th, 8th and 9th, to the 10th class; the 11th to the 13th; and the 18th, 21st, 22d and 23d, he included in one, and thus reduced the number of classes to 16. He has had few followers; therefore, we shall not stop to point out the impropriety of some of these combinations.

Thunberg has reduced the number to twenty, by distributing the plants of the 20th, 21st, 22d and 23d classes, among those classes to which they are referable, according to the number and connexion of the stamina. His reasons for abolishing the classes Monoecia, Diœcia and Polygamia, are, 1st, The plants which belong to them are not always constant in point of sex; for culture, and a difference of climate, frequently removes a plant from the Monœcia, or Diœcia, to the Polygamia. 2dly, By arranging the genera which compose these classes according to the nature and connexion of their stamina, many of them will stand in the same class with others which they resemble in almost every particular, except in the circumstance of having their male and female organs placed in separate flowers, either on the same or on different plants. His reasons for abolishing the class Gynandria are not equally valid:

Gmelin, in addition to the alterations proposed by Thunberg, which he has adopted in his edition of the Syst. Nat., has likewise united the class Icosandria to Polygamia, and consequently reduced the number to nineteen. This union is certainly unnecessary, if not altogether improper, as these two classes are sufficiently distinguished by the insertion of their stamina; a distinction which he himself has retained in his subdivisions of the united classes.

Willdenow, we think, has done right in rejecting all these proposed reductions in the number of classes; for it rarely happens that any considerable alteration can be made in a long established system, without producing some confusion and inconvenience therefore none ought to be made or adopted, but such as evidently tend to produce some very considerable improvements, which is not the case with any of those hitherto made on the sexual system. The only deviations he has made from the arrangement of Linnæus, are, first, to break up the order Monogynia, of the class Syngynesia, and to place the plants which composed it, according to the number of the stamina, without any regard to their connexion by the antheræ. The plants belonging to this order, certainly have no affinity to the others which compose the class Syngynesia, which becomes a perfectly natural assemblage after their removal. Secondly, He has removed a great number