are narrower and more acute than in the ordinary Chilian form, but are in this respect undistinguishable from some collected by Gillies; the carpels of the two are identical. The G. elatum, of Kamaon, judging from Dr. Wallich's specimens, does not at first sight appear to differ from this, except in the comparatively smaller terminal and the sessile lateral lobes of the leaf. Dr. Royle has placed it in Sieversia, to which genus Mr. Edgeworth assures me that it belongs. The nearest, but evidently distinct, species allied to G. Magellanicum are G. Pyrenaicum, easily recognised by the great size of its carpels, and G. sylvaticum of the South of France and Spain, which is a single-flowered plant.
2. Geum parviflorum, Commerson ; velutino-pubescens, rliizomate crasso, foliis radicalibus interrapte pinnniatisectis lobo terminali rotundato obscure 5-lobato crenato lateralibus 2-3-jugis multoties minoribus, pedunculis folio brevioribus elongatisve pubescentibus folia 2-3 lyrato-pimiatifida gerentibus, floribus 4-5 ad apicem pedunculi sessilibus nutantibus folio involucratis, petalis lacixdis calycinis subsæquantibus albis? ovarii stylo hamato, carpellis pilosis. G. ? parviflorum, Commerson ex Smith in Rees Cycl. vol. v. p. 16. DC. Prodr. vol. ii. p. 553. Don, Encycl. vol. ii. p. 527. G. involucratum, Juss. Herb, in Pers. Each. vol. ii. p. 57. DC. et Don, 1. c.
Hab. Strait of Magalhaens ; Commerson. Port Famine; Capt. King.
Rhizoma pollicem crassum. Folia fere omnia radicalia, patentia, 1½—2 unc longa, pilis fulvis dense velutina, lobo terminali ¾ unc. lato, plicato, lateralibus irregulariter ineiso-dentatis. Pedunculi 3, unico foliis breviore, cæeteris elongatis erectis ter longioribus, omnes basi nudi, apices versus folia 1-2 gerentes. Flores in capitulum aggregati, sub ¼ unc. diametro, albi (fid. Commerson). Calycis laciniæ 6, oblongæ, obtusæ, bracteolis lineari-oblongis. Petala late elliptico-spathulata, obtusa, glaberrima. Ovaria dense hirsuta, stylo curvato apice hamato terminata.
A little known and probably very rare plant. I have seen but one specimen, in Capt. King's collection, winch I have compared with that in the Linnsean Herbarium; its general appearance resembles a small state of G. Magellanicum, but the leaves are densely velvetty on the surface and the flowers very different; Commerson says the latter are white, in which respect, as in their size, the form of the petals, &c., there is a close affinity with the Sieversia? albiflora (vol. i. p. 9. t. vii.), a plant which may possibly in an older state have hooked awns to the carpels, which these decidedly are. I am inclined to consider this the representative of the Auckland Island species, as the G. Magellanicum is of a similar New Zealand one. Though the descriptions of G. involucratum are very unsatisfactory, I have little doubt but that they refer to this plant as above described, which seems to have been first published in France and afterwards in England. The description of Smith being the fullest, and his having adopted Commerson's own name induce me to retain that of G. parviflorum. In the flower I examined there were six divisions to the calyx.
3. RUBUS, Linn.
1. Rubus geoides; Smith, Icon. ined. t. 19. Hook. Icon. Plant, t. 495. R. antarcticus, Banks et Solander, MSS. et Ieon. in Mus. Brit. Dalibarda geoides, Pen. Euch. vol. ii. p. 53. BC. Prodr. vol. ii. p. 568. Gaud, in Ann. So. Nat. vol. iv. p. 106. et in Freye. Voy. Bot. p. 138. B'Urrille, in Mém. Soc. Linn. Paris, vol. iv. p. 620. Framboise, Pernetty, Voy. vol. ii. p. 58.
Hab. Strait of Magalhaens; Commerson. Port Famine; Capt. King. Good Success Bay; Banks and Solander. Falkland Islands; abundant on the bills.
There are few Rosaceæ in the temperate latitudes of the Southern Hemisphere. The genus Rubus, in particular, hardly exists in South America; a very remarkable circumstance, since Australia, New Zealand, and the Cape of Good Hope possess more species than extra-tropical South America, though those countries are more dis-
