ON THE PROTEACE^E OF JUSSIEU. 141
tissime marginatis et apice brevissime alatis : vel lignc'i, subrotundi, pseudo-bivalves bast tantum styli mucronati ; seminibus undique alatis.
This extensive genus, of which a few of the least remarkable species have been already published as Em- bothriums by Dr. Smith, Cavanilles, and others, I have dedicated to the Right Honourable Charles Francis Greville, one of the Vice-Presidents of the Royal Society ; a gentleman eminently distinguished for his acquirements in natural history, and to whom the botanists of this country are indebted for the introduction and successful cultivation of many rare and interesting plants.
Grevillea is probably the most extensive genus of Pro- teacese in New Holland, and admits of division into several very natural sections, most of which are readily distin- guishable by more than one character, existing either in the parts of fructification or in habit; notwithstanding which, I have not ventured to separate them into distinct genera, as I probably should have done, had I been ac- quainted with fewer species ; but have given to each section a proper name, a practice that may perhaps be advan- tageously adopted in all large genera where they are thus capable of natural subdivision. It must be unnecessary to add that proper names can in this manner be given only where the sections are perfectly natural, and not in those cases where genera have been subdivided from single characters, and those too of but little importance, as in Thunberg's division of Protea, from the form and division of the leaves ; to which may be opposed the masterly ri69 subdivision of the same genus previously given by Linnaeus in the Mantissa, whose sections, though apparently depend- ing on single characters, are evidently formed from a con- templation of the whole structure, as far as it was then understood ; and it is remarkable that, with the exception of the first species, with whose real structure he was neces- sarily unacquainted, the rest are arranged, and even divided into sections, in most cases corresponding with the genera proposed in the present essay.
�� �
