NATURAL CLASSIFICATION Or PLANTS.
posed according to their technical characters. Motu 'ray, in compiling the fourteenth edition of that work, has been inadvertent, respecting this essential part of its plan. Indeed it is probable that he was not competent to judge of the affinities of the new genera, introduced from the Suppkmentunt, or from the communications of Jacquin, Thunberg, &c. Yet surely he might have perceived the affinity of Banksia to Protea, rather than to Ludwigia or Oldenlandia ; and indeed Linnaeus himself ought to have discovered the relationship of the latter to Hedyotis, if he did not detect their identity, instead of inserting it be tween two such strict allies of each other as Ludwi gin and Ammannia. To pursue these remarks would be endless. It is hardly necessary to indicate the natural classes. or orders, of the Linmean system, such as the' Tetradynamia, Didynamia, Diadelphia, Syngenesia ; the 7'riandria Digynia, Gynandria Diandria, &c. Except the first-mentioned class, which, if Ckonte be removed, is strictly natural and entire, the others are liable to much criticism. We are almost disposed to allow, what we know not that any one has yet observed, ;hat the system in ques tion is the more faulty in theory, for these classes being so natural as they are. Each order of the Didynamia presents itself as a natural order, though the character of that class, derived from the propor tion of the stamens, serves to exclude several genera of each order, and to send them far back, into the second class. If all ideas of natural affinity be dis carded from our minds, there is no harm whatever- in this ; but if the Didynamia claims any credit, as a class founded in nature, the above anomaly is a de fect. So, still more, under the same point of view, is the Diade4thia, or at least its principal order De. candria, liable to exception. This order eonsists entirely of the very natural family of Papilionacete. They are characterized as having the ten stamens in two sets. Now it happens that there are many papilionaceous genera, indeed a great number of such have been discovered since Linnaeus wrote, whose ten stamens are all perfectly distinct. These therefore are necessarily referred to the class De candria, and they come not altogether amiss there, because they meet, in that class, some concomitant genera, which though, like them, leguminous, are less exactly, or scarcely at all, papilionaceons. But
the greatest complaint lies against some genera of the Diadelphia Decandria, for having the stamens all really combined into one set, so as in truth to answer to the technical character of the preceding class, Monadelphia. There is mostly indeed some indication of a disunion upward, where they, more or less perfectly, form two sets ; and some of them are so nearly diadelphous, that their complete union at the bottom may easily be overlooked ; others, however, have only a fissure along the upper side of their common tube, without any traces of a separate stamen or stamens. The papilionaceous character of the corolla therefore, in such cases, is made to overrule that of the particular mode of union among the stamens, and is in itself so clear, as seldom to be attended with any difficulty ; but the incorrect ness of principle in the system, in the point before us, es being neither professedly natural, nor exactly artificial, cannot be concesled. Part of the objec tions, to ,which the sexual system was originally liable, have been obviated. We mean what concerns the last class but one, Polygatnia. Dr Forster ob served, in his voyage round the world, that this class was subject to great exception, on account of the trees of tropical climates, so many of which are constantly or occasionally polygamous ; that is, each individual frequently bears some imperfect flowers, male or female, along with its perfect or united ones. Such a circumstance reduces any genus to the class Polygantia ; and on this principle Mr Hudson, thinking perhaps that he made a great improvement, removed our Iles Aquifolium,. or Holly, thither, though Ike is well placed by Linnaeus in the fourth class. The author of the present essay has ventured to propose a scheme, which is adopted in his Flora Britannica, for getting clear of this difficulty. He considers as polygamous such genera only as, be sides having that character in their organs of im pregnation, have a difference of structure in the other parts of their two kinds of flowers. Thus Atriplex has, in its perfect flowers, a regular spread ing calyx, in five equal segments ; in the attendant female ones a compressed one, of two leaves, subse quently much enlarged.