same and the properties, only varying in degree, also the same, no one surely will deny that such genera are unworthy of science and altogether misplaced in any system professing to aspire to the character of natural.
It is surely time we were bidding adieu to such puerilities and studying, not how far we can split and multiply genera by restricting our characters within the narrowest limits, but how we may so construct them as to include every species that naturally belong to them, and to exclude all that do not, a point of perfection which, I fear, we shall not soon attain, so long as we use the varying shapes and sizes and duration of deciduous parts as generic characters and talk of natural systems, but deny that either natural orders or natural genera come from the hand of nature.
Influenced by these considerations, I reject all characters taken from the mere external form of the calyx, whether nearly truncated or lobed, long or short, also whether the petals are so caducous that they fall before expansion or fairly expand and prove as persistent as petals usually are in tropical climates, that is, have a duration of from 12 to 24 hours. Such characters applied to the distinction of genera, it appears to me, are well fitted to establish the truth of the axiom that nature does not create genera, but at the same time indicates much want of philosophy in our manner of interrogating nature and enquiring for natural genera, since such distinctions can only produce the most artificial combinations of species.
But discarding them, and looking only to structure, not o size or relative permanency of deciduous parts, we can approach more nearly to the construction of natural genera. The genus Myrtus might then be confined to such species as have quinary flowers, a 3-celled ovary, baccate fruit, and several bony seed enclosing a somewhat cylindrical embryo. Myrcia to those having quinary flowers, a 2-celled ovary, seed with a smooth not bony testa, and foliaceous corrugated cotyledons. Jossinia to those having quaternary flowers, a 2-celled ovary, numerous seeded fruit, like Psidium or Myrtus, and foliaceous cotyledons.
Thus defined, these appear to be all good genera, but, not having materials to compare, I can offer no decisive opinion on that point, though think it very desirable to ascertain whether a genuine Myrtus ever has a quaternary flower and 2-celled ovary, or a Myrcia a quinary flower and 3-celled ovary, for I think not : or would it not be better to class the species of these genera according to characters taken altogether from the flower and ovary, even should they occasionally be found somewhat arbitrary, because, as now defined by DeCandolle, it is impossible to distinguish a Myrtus from a Myrcia without ripe seed, which, for practical purposes, is nearly useless, we so seldom find fruit in that state in herbaria.
The genera Caryophyllus, Eugenia, Jambosa, Syzygium, and, I believe, Acmena D.C. all associate in having a 2-celled ovary (ever 3 ?) with numerous ovules attached to the inner angles of the cells, subdrupacious fruit, with few, generally one or two, globose fleshy seed, variously divided. By subdrupacious, I here mean a fruit consisting of an exterior pulpy or fleshy portion, sarcocarp, enclosing the seed, in this case not hard and bony, as in true drupes, but fleshy and easily sectile. The testa of the seed is besides so very thin, that unless looked for, it may be overlooked. The cotyledons or body of the seed is quite peculiar, altogether sui generis. In all, the radicle is small and inconspicuous with comparatively large seed lobes, these lobes being sometimes conferruminate, that is adhering by their margins so as to appear one only, at others, divided into two or three or a dozen lobes but all united in the minute central radicle.
The insufficiency of lobed or conferruminate seed to form a generic character is proved by the fact of both forms being found on the same plant. In these structural peculiarities, which pervade the whole tribe of Eugeniae we have, it appears to me, conclusive evidence that nature does create genera and that this group, which presents nearly every variation of vegetable form and inflorescence, being yet pervaded, through its whole extent, by a uniform structure in the organs most essential to the preservation of the species, shows that it is truly one of nature's own genera and, as such, ought on no account to be broken down and frittered away by the introduction of frivolous distinctions without practical value or facility of application when employed in practice, since in their nature they are fluctuating and unstable.
The oldest name of the group of genera, which I propose uniting into one, is Caryophyllus, and under that name the whole phalanx should be ranged, with the essential character — flowers quaternary, ovary 2-celled, with numerous ovules attached to a central placenta, seed thick and
