partaking more of the character of Memecylon than of Myrtus. Turning next to Eugenia, which in its more obvious characters, those namely taken from the calyx and flower, does not differ materially from Myrtus but is separated by its 2-celled ovary and very different seed. These, which in Myrtus, are compressed reniform enclosing a terete curved embryo in a hard bony testa, are, in Eugenia globose, with thick soft and fleshy cotyledons ; forming an excellent distinction between these often, otherwise, closely allied genera.
When we next proceed to compare Eugenia with its off sets, Jambosa and Syzygium, permanent distinctions are not so easily found. All three have a 2-celled ovarium with numerous ovules, all have a more or less succulent fruit, all have fleshy seed. The distinctions, therefore, attempted to be established rest on Jambosa having a turbinate somewhat elongated calyx tube, a grumous edible fruit and a lobed seed, in place of a globose calyx tube and conferruminate seed. The first of these can scarcely be admitted if Wallich's E. formosa and E. bifaria are to be considered genuine species, since both have the turbinate calyx and general habit of Jambosa and, in my opinion, are both more justly referable to that genus, as it now stands, than to Eugenia. The character taken from the cotyledons being lobed or conferruminate, does not appear a more satisfactory one and must, as in the instance just quoted, be received with caution, as both seem to have conferruminate cotyledons though that is not mentioned in the description. Syzygium is said to be distinguished from both, by the petals cohering, forming a calyptra or lid and falling off either in that state or immediately after expansion, thus making the essential character of the genus rest on its having caducous petals ; the seed being the same as in Jambosa. This seems to me at best a very artificial character and, if not taken along with habit, is altogether nullified by several species in my collection, which have the habit of the most perfect forms of the genus, but want the deciduous petals. To establish generic characters on such distinctions, which are barely sufficient to supply very secondary sub-divisions, is altogether to banish from natui'al history the axiom that — the genus gives the character not the character the genus.
When we extend our comparison a step further and compare the distinctions taken from the calyx tube in these two genera Syzygium and Jambosa, we find them equally at fault. Syzygium is said to have the calyx tube obovate, while in Jambosa it is turbinate and attenuated at the base. In S. Zeylanicum D.C. and S.WightianumWall. it is described as elongated and clavate, while in both, the inflorescence is racemose, the flowers springing from the naked branches, as in Jambosa malaccensis and some others, in place of forming umbelliform cymes, the usual form in the genus, thus corresponding in habit as well as structure with Jambosa. For these reasons I cannot consent to separate Syzygium and Jambosa. Eugenia it seems to me must equally be united, if we would form a natural association.
Let us next consider Caryophyllus, the oldest genus of the group. It is characterized by having a cylindrical calyx tube and 4-lobed border, the petals cohering into a lid, as in Syzygium, and the stamens free, forming 4 tufts with a slight almost inconspicuous interval between : the inflorescence is corymbose. Here again we have the identical structure mentioned as belonging to S. Zeylanicum and S. Wightianum, with the exception of the 4-lobed limb of the calyx and 4 tufts of stamens, the fruit and seed are much the same in all — but surely the limb of the calyx being a little more or less divided can never be received as a generic character being, in truth, only fit to be employed as a specific one. For these reasons I propose to reunite these four genera Eugenia, Caryophyllus, Jambosa and Syzygium into one, and use those characters which have hitherto been employed as generic ones, for its subdivision into sections, it being quite impossible they can ever furnish good generic ones : genera which are made to depend on such characters as a greater or less development of any single organ, such as in the elongated calyx of Caryophyllus and of Jambosa, in contradistinction to the short one of Syzygium, or on the more or less deciduous petals of species, cannot be natural. Neither can we get good artificial genera, by the combination of these characters, unless we are prepared to multiply them beyond measure. In that case it might be easy enough to construct generic characters — for one genus might have a short calyx tube and petals that expand before falling — another might have an elongated tube and similar petals — a third might have a long tube with deciduous calyptrated petals — a fourth a short one, and so on and if rigidly adhered to, might facilitate the determination of species, but while the ovary, ovules, and fruit are nearly the same in all, the general habit the
