number of stamens, far above twenty; such as Poppy, Capparis, Cistus, Ranunculus, and others. But in the class Icosandria, there are not a few genera which are as abundantly supplied with these sexual organs, such as Cactus, Eugenia, Rosa, and others. By attending, however, to the very different manner of insertion of the many-stamened plants of these two classes, we shall have no difficulty in referring them to their proper classes.

Be

PROFESSOR Gmelin, in his edition of the Systema Natura, has united the two classes Icosandria and Polyandria into one, which retains the name of POLYANDRIA! I cannot but consider this as a very injudicious alteration; and I am happy to unite in sentiment, on this subject, with a number of able botanists *. The classes Icosandria and Polyandria should be kept apart, and not merely in subordinate divisions, but in the higher associations, or classes. Nature is remarkably regular and constant in the place of insertion of the stamens. sides, by uniting, or rather confounding, the two classes in question, we render the sexual system much more artificial than it really was, when it came from the hands of its GREAT ARCHITECT: we, thus, deprive it of one of its charms, its occasional approach to a NATURAL SYSTEM. For although the system of Linnæus is professedly artificial, it has, certainly, some claim to the character of a natural assemblage of vegetables; and I cannot help thinking, that in proportion as it is render. ed more artificial, it will lose a part of its value.

It has already been observed, that the class Icosandria contains a great number of esculent and innocent vege

* Professor Martyn, Dr. J. E. Smith, Dr. Darwin, Professor Willdenow, and others.

tables. We shall soon see, on the other hand, that the class Polyandria abounds in poisonous or active vegetables. In this respect, the two assemblages are very different from each other, and this circumstance should have had some weight with rash innovators, before they made the disposition to which I have objected. In fact, the mode of insertion of the stamens in the class Icosandria is a feature of great importance. Even in other classes, a like mode of insertion sometimes gives an indication of the wholesome properties of the vegetable. This is the case with the genus Ribes, comprehending the different kinds of Currants, Gooseberries, &c. which are some of the most innocent fruits in the class Pen-' tandria*.

WHILST, however, I object to the melting down of the two classes Icosandria and Polyandria into one class, I confess, again, that I should have no objection to see the last of these classes, in some measure, blended with the class Dodecandria: that is, I think it would be well to introduce all the true Dodecandrous plants, having more than ten stamens, into Polyandria. This would, certainly, facilitate the labour of the student. For, after being told, that the Dodecandrous plants have from twelve to nineteen stamens inclusive, with what propriety do we include in the class Polyandria a number of genera in which no botanist perhaps, has ever observed as many as nineteen of the male sexual organs: in which, at least, the number nineteen is exceedingly In Podophyllum, Sanguinaria, and some other

rare.

* The stamens of Ribes are inserted into the calyx.

genera, we rarely observe more than fifteen or sixteen stamens *.

THE class Polyandria is subdivided into seven orders, viz. MONOGYNIA, DIGYNIA, TRIGYNIA, TETRAGYNIA, PENTAGYNIA, HEXAGYNIA, and POLY

GYNIA.

MONOGYNIA.

ONE FEMALE.

THIS first order contains the following genera: viz. Marcgravia, Ternstromia, Alstonia, Trilix, Rheedia, Mammea, Papaver, Cheledonium, Capparis, Actea, Cambogia, Calophyllum, Grias, Sparmannia, Mentzelia, Sloanea, Cistus, Corchorus, Sarracenia, Tilia, Ochna, Muntingia, Elæocarpus, Myristica, Argemone, Lagerstroemia, Thea, Lecythis, Sanguinaria, Jeffersonia†, Podophyllum, Bixa, Nymphæa, Trewia, Laetia, Seguieria, and Delima.

* In some of the earlier editions of the Genera Plantarum, several genera which were afterwards introduced into Dodecandria, had actually a place in Polyandria. I here mention the names of these genera, viz. Bocconia, Cratæva, Euphorbia, Peganum, Portulaca, Triumfetta, Heliocarpus, and Reseda. See Genera Plantarum, &c. Editio secunda. Lugduni Batavorum: 1742.- See Part I. page 68. note.

Jeffersonia has, very generally, only eight stamens. Sometimes, however, it has more; but never, I believe, as many as twenty. From its near affinity to Podophyllum and Sanguinaria, it may with as much propriety be introduced into the class Polyandria, as these two genera.

DIGYNIA.

TWO FEMALES.

THIS order is much less extensive than the first. It contains the following genera, viz. Calligonum, (which is referred by Willdenow to the class Dodecandria), Fothergilla, Curatella, and Pæonia.

TRIGYNIA.

THREE FEMALES.

THE genera Delphinium and Aconitum are referred to this order; as is, also, a species of Reseda, Reseda Luteola, called Yellow-weed, or Dyers-weed.

TETRAGYNIA.

FOUR FEMALES.

THIS order contains the genera Tetracera, Caryocar, and Cimicifuga.

PENTAGYNIA.

FIVE FEMALES.

AQUILEGIA, Nigella, Reaumuria, and Brathys be

long to this order.

HEXAGYNIA.

SIX FEMALES.

THE genus Stratiotes, or Water-Soldier, and Brasenia belong to this order.

POLYGYNIA.

MANY FEMALES.

THIS beautiful order contains the following genera, viz. Dillenia, Illicium, Liriodendron, Magnolia, Michelia, Uvaria, Annona, Nelumbium, Anemone, Atragene, Clematis, Thalictrum, Adonis, Ranunculus, Trollius, Isopyrum, Helleborus, Caltha, Houttuynia, Hydrastis, Drymis, and Unona. Also a new and singular NorthAmerican genus, of which two species are now known. They are aquatic plants, with peltate leaves.

CHARACTER.

By some writers, the class Polyandria has been deemed a natural class. As such I cannot consider it, though it is, unquestionably, more natural than some of the other classes of the sexual method. It is one of those classes which comprehend several pretty natural families of vegetables, that are related to each other by affinities

I