their real affinities, il ; becomes a curious and interesting, though by no means easy enquiry, to trace the various attempts that have been made by different authors, to construct a natural arrangement of these plants while excluding this their only truly natural and constant bond of union. Some of these arrangements I shall endeavour to exhibit, partly in support of (he assertion advanced, that no two Botanists are agreed on these affinities, but principally to prove that, so long as we take the different degrees of development of the flower as the basis of our natural system of Botany, we build on a bad and unstable foundation.
Lindley has employed this character, in the construction of his Alliance Silenales, the per- fection of which, however, through his not perceiving the full value of his character, "embryo rolled round mealy albumen," he has marred, by the introduction of Tnmariscinene (an order in which it is wanting and hut remotely if at all allied) on the one side and the exclusion of Ficoideae on the other. This last he has in my opinion most injudiciously placed in his alliance Cucurbitales, for no reason that I can perceive, its fruit being neither epigynous nor its placentas parietal.
The following are the polypetalous orders which participate in the character and ought to have been associated in that alliance. I use Lindley's names. Portalace.ae, Silenaceae, Alcina- ceae, lllecebraceae, Ficoidaceae, and perhaps Surianaceae. The remaining orders agreeing in this character? in number are arranged by Lindley in his group Curvembryosae ( sub-class incomplete ) they are, Amaraniaceae, Chenopodiaceae , Phytolaccaceae, Polygoneae, Petiveraceae, Scleran- thaceae and JYyctaginaceae. Menispermaceae, placed in this group on account of its curved em- bryo, I exclude, as having no other affinity, its albumen not being mealy nor the embryo on one side curved round it. This it will be perceived is considerably different from DeCandolle's ar- rangement who places Caryophylleae in his first class, Thalamip'orae — Portu aceae, Paronychiaceae, and Ficoideae in his second — Calyci/iorae — and the remaining order in his fourth class Monochlamydeae.
Von Martius adopts a different arrangement but being of a mixed character cannot be so easily explained, but the following list of the orders, numbered as they stand in his Conspectus, will show how widely the group is scattered over his system — 90 Chenopodiaceae — 91 Rii^in iaceae • — 92 Petiveraceae — 93 JVyctagineae — 94 Sclerant/ieae — 95 Phytolacceae — 96 Polygoneae — 235 Amaraniaceae — 236 Paronychieae — (a) lllecebieae — (b) Polycarpeae—(c) Minunrtieae — 261 Portulaceae —262 Caryophyllnceae — (a)Sileneae — (b) Alsineae f 263 — Elalineae. The absence of albumen in the last seems to exclude it from the group. In Endlicher's Genera Plan- ter Chenopodeae ranks No. 101 —Amarantaceae. 102 — Polygoneae 103 — JVyctagineae 104 — Mesembryan t hemeae or Ficoideae 205 — Portulaceae — Caryophylleae ^07. (The subse- quent parts of the work have not yet reached me) — The first four of these orders form his class Oleracene — the last three form part of his class Caryophyllinee which includes all the genera in our Prodromus referred to Caryophylleae, Portulaceae, Ficoideae and Pa> • oni/chiaceae. Under the name of Tetragoniaceae Lindley refers Sesuvium and Aizoon to Endlicher's Oleracene placing them between Amaranthaceae and Chenopodiaceae on the one side and Phytolacceae and Polygoneae on the other.
Could we ask for more conclusive evidence of the intimate relationship existing among all the orders and genera having that very peculiar structure of the seed, than the facts here stated supply ? No sooner do we find one Botanist, eminent for his knowledge of affinities of plants, distribute the orders so marked, in the way he thinks most consonant with their natural affinities, than we find another equally celebrated proposing a different arrangement. DeCandolle placed so many of these genera (those of his order Caryophylleae), in his class Thalamifloreae ; so many more (the orders above named) in Calyci florae ; and the remainder, those referable to Chenopodiaceae, Amarantaceae, &c, in his Monochlamydeae ; thus distributing them all over bis system. Lindley, in the first, edition of his natural system of Botany, arranges no fewer than 165 orders in a single series, under the heading " Polypelalous Apetalous and Achlamydeoas Plants," altogether unlramelled by system or arbitrary divisions of any sort, and here, it is remarkable, that of 11 orders possessing this structure — mealy albumen with the embryo on one side — 9 are, with two exceptions, placed in immediate succession, Nitrariaceae standing between Ficoideae and Illecebreae, and Begoniacene, between Poly galeae, and Nyctagineae, The remaining two Caryophylleae, and Portulaceae, are near.
This circumstance shows (hat their general affinities are such, that the whole might, by that single character, be separated from the rest of the system and grouped in one very natural
