too great violence to Nature, and swerving from that beautiful and philosophical Linnæan principle, of characterizing genera by the fructification alone; a principle which, those who are competent to the subject at all, will, I believe, never find to fail. The seeds and flowers of the umbelliferous family are quite sufficient for our purpose, while the involucrum is very precarious and changeable; often deficient, often immoderately luxuriant, in the same genus. In the cymose plants every body knows the real parts of fructification to be abundantly adequate, the involucrum being of small moment; witness that most natural genus Cornus. For all these, and other reasons, to particularize which would lead me too far, I have, p. 236, reckoned the Umbel and Cyme modes of flowering, and not themselves aggregate flowers.
