envelopes, we should have classed the latter among the essential organs of the first class.
The importance of characters, in as far as derived from the importance of the organs they relate to, would follow the same gradation,—observation (not theory) teaching us, however, to place those derived from the reproductive organs of each degree before the corresponding ones derived from the nutritive organs; and those derived from the embryo or young plant, more especially at the moment of germination, above all.
But the second element in the ratio of value of characters, the point of view in which the organs are considered, is one which experience shows to be often far more important than the nature of the organ itself, and the neglect of which contributes more than anything to the degeneracy of an apparently natural classification into a purely artificial one. The principal characters which an organ, or set of organs, can thus supply, and their relative importance, are admirably expounded by De Candolle, in his Taxonomie, div. I., chap. 3. He there establishes the following scale of gradation, in which I have ventured to make some slight modification in expression, but which I think should never be lost sight of by the systematist who has any pretension to establish natural groups.
1. The real presence or absence of organs (parts of organs, or sets of organs), independent of adherence or accidental abortion.
2. Their arrangement, or relative position, and numbers, as affecting or indicating the general plan upon which the plant is constructed.
3. Their external form, relative size, continuity or articulation, &c., all subordinate to the preceding class, only acquiring importance when indicative of a result from general arrangement.
4. Their functions and sensible qualities,—the results, rather than the causes, of the preceding modifications.
By combining this scale of relative importance with that derived from the nature of the organs themselves, it might be possible to frame a general scale of relative importance of characters, which, with other rules suggested by the observation of the comparative prevalence of particular characters, might assist in judging of the expediency of describing as a new genus or order any newly-discovered plant which does not come precisely within the limits previously fixed for any known genus or order. But, in the grouping together any number of species or genera already known, the relative value of the characters relied upon should be tested, at every step, by a comparison with all the other features of the plants. The blind adherence to a pre-established scale, in distributing into genera the species of a large order, renders such a classification purely artificial. It had been ascertained that the relative arrangement of the radicle and cotyledons in the embryo of Cruciferæ, the relative prominence of the ribs of the fruit, and the number and arrangement of the vittas in Umbelliferæ, the various modifications of the pappus in Compositæ, were in many cases remarkably constant, not only in species, but in many very natural genera. But by taking these characters as absolute, and considering every slight modification of them
