oblongo-ovatis, valde concavis, acutis v. subito acuminatis, medio carnosis, marginibus late subscarioso-membranaceis. Corollæ tubus latiusculus; limbus quadrifidus, segmentis ovatis, acutis, marginibus involutis, medio late uninerviis. Staminum filamenta basi lata; antheræ oblongo-sagittatæ. Pollen globosum, flavum. Ovarium parvum, elliptico-ovatum, compressum, utrinque sulcatum, spurie biloculare: ovulis 6—8, peltatis, placentæ columnari affixis. Stylus simplex, erectus, filiformis, per totam longitudinem glanduloso-pilosus, apice simplex. Capsula chartacea, brunnea, late obovata, turgida, paulo compressa, basi stylo apiculata, infra medium circumscissa. Columna centralis libera, erecta, clavata, bialata, alis ultra columnam in cornubus 2 productis. Semina valde depressa, 6—8, etsi compressione mutua angulata, pleraque triangularia, pallide testacea, translucida; testa mucilaginosa, membranacea; albumen inter corneum et carnosum, semini conforme; embryo compressus, ut mihi videtur oblique transversus; radicula obtusa, ad hilum paulo versa; cotyledones majusculæ, plano-convexæ.
This plant was first detected by Mr. Brown in Tasmania, in the southern parts of which island it is not uncommon, growing, as in Lord Auckland's group, on maritime rocks, and from whence we have specimens of both varieties from Mr. Gunn. In habit and foliage it most resembles the P. Coronopus of any European species, but it differs totally from that plant in the structure of the capsule, as also in the inflorescence. It belongs to a small group of the genus, not hitherto recognised, but which are remarkable for having the flowers solitary or truly capitate and not spiked, never more than from three to five in number. This inflorescence differs very much from that of several species in which the spike is abbreviated, either naturally or by accident, to a few flowers.
The species naturally allied to P. carnosa are the P. rigida and nubigena, H.B.K., P. monanthos, D'Urv., P. andicola, Gill., P. pauciflora, Lam., and P. barbata, Forst., all natives of the southern regions of the globe and of the western hemisphere, whose only representative in the old world is the present plant. If however the sections proposed by Endlicher are adopted, these species will be found to be in several cases widely separated from one another. P. nubigena, a very beautiful species, has only two-seeded capsules; the seeds are quite unlike those of carnosa, being large, elliptical-oblong, black and punctated, with a broad hollow on the face. Of P. rigida we have specimens gathered by Mr. Mathews on the Pampas of the Cordillera in Peru; it is perhaps the most singular species of the genus in the structure of its flowers, which are solitary, almost sessile, on very short peduncles, surrounded by a very broad sheath or spatha rather than bract at the base; the calycine segments are lanceolate, acuminate, with a tuft of silky hairs at the base; the tube of the corolla is twice as long as the calyx, slender, and at least three times longer in proportion to its breadth than in any of the genus which I have examined; the filaments very long, straight and erect. M. Kunth places it in a section "capsular loculis 1-spermis," but says he has not seen the fruit. From the small size of the ovarium I could not detect the ovules in the flower I examined. P. monanthos, though generally single-flowered, has often two or even three flowers; the capsule is four-seeded; it is a very distinct species, common in the Falkland Islands and in Fuegia, singularly variable in size, and in favourable situations often becoming spuriously caulescent, with stems similar to those of the P. arborescens of Madeira; the leaves are then two inches long; whilst in other situations, as on the exposed rocks of Cape Horn, the whole plant resembles a densely tufted moss, the leaves being rosulate and not two lines long. Had I not the opportunity of examining it in its native state, I should not have considered several of its varieties as belonging to one and the same species. P. andicola, Gill., is a very singular plant with roots as thick and long as the little finger, the collum often much thicker, bearing two or more capitula of very coriaceous, lanceolate, smooth or pilose leaves, about 1½ inch long. The scapes are longer, often twice as long as the leaves, slender and hairy, with three to five flowers. The ovarium contains several ovules, and the capsule four or more seeds, rarely less, but sometimes only two. It has been gathered on the Andes of Chili, both by Dr. Gillies and by Mr. Bridges. P. pauciflora, Lam., and P. barbata, Forst., seem involved in some confusion; I have seen no specimens answering to the description of either, if they be truly distinct. In Mr. Anderson's collection, formed during Captain King's voyage, there is a species (P. imberbis, MSS.) agreeing with
