and observations show tbat A. alpina of Weber and Mohr (Bot. Tamil, t. 11. f. 3, 5), is only a state of A. rupestris, with leaves spreading in all directions. This is confirmed by the fact that A. alpina of Mougeot and Nestler (Stirp. Crypt. Voges. no. 115), is that very form of A. rupestris to which we allude. The figures of Dillenius doubtless refer to A. alpina ; but the absence of a separate figure of A. rupestris, and the remark, "in rupibus surculi e fusco rufescunt" (p. 507), prove that this author considered both these as varieties of one species. The true A. alpina may always be known from A. rupestris by its obovate dark glossy leaves, and its longer and more robust stems.
2. Andreæa marginata, Hook. fil. et Wils.; caulibus laxe csespitosis subramosis, foliis erecto-patentibus incurvis ovatis longe acuminatis infra medium coutractis enerviis marginibus inferne pallidis caulinis majoribus superioribus confertis, periehaetialibus longioribus elongatis late lanceolatis convolutis, theca exserta. (Tab. CLI. Fig. I.)
Hab. Hermite Island, Cape Horn, frequent on rocks on tbe mountains ; J. I). II.
Caules unciales. Folia e basi ovata longe acuminata, concava, superiora in caule fertili sensim majora, laxiora et magis erecta, omnia puniceo-atra, nitentia, areolis minutis, oblongis, marginRbbus inferne minoribus, pallidioribus.
With some hesitation we venture to separate this from A. alpina, on account of the more gradual acumination of its leaves, which are full twice as long. It differs from A. acutifolia in having the leaves considerably larger, the lower part more suddenly dilated, and the base not gibbous; also in their dark colour and glossiness, and in the more evident perichffitium. It is easily recognized by its habit.
Plate CLI. Fig. I. — 1, a specimen : — natural size; Jig. 2 and 3, leaves; fig. 4, theca : — magnified.
3. Andre/EA acutifolia, Hook. fil. et Wils., vid. Part 1. p. 118. Var. fi. rufescens, ramis fastigiatis. (Tab. CLI. Fig. II. 2.)
Yar. y. foliis latioribus.
Var. 8. foliis superioribus subsecuudis. (Tab. CLI. Fig. II. 1.)
Var. c foliis superioribus ercctis elongatis.
Var. f. foliis erectis obtusiusculis atro-sanguineis.
Hab. Var. a, 0, and e, Hermite Island, Cape Horn. Var. 8 and (, Falkland Islands. Var. y, Kerguelen's Land.
All these varieties differ somewhat, though slightly, from that gathered in Campbell's Island.
Plate CLI. Fig. II. — 1, specimen of var. 8, and 2, specimen of var. /3: — natural size; fig. 3, perichaetium and theca ; fig. 4, perichostial leaves ; fig. 5, cauline leaf of var. 8 ; fig. 6, leaf, and fig. 7, theca of var. : — magnified
4. AndeejEa rupestris, Linn.; caule humili subramoso, foliis e basi vaginante patentibus (interdum seeundis) ovato-lanceolatis vel ovatis superne attenuatis acutiusculis enervibus superioribus siccitate appressis, pericbsetialibus longioribus ovato-lanceolatis convolutis, theca exserta. A. rupestris, Hook, et Tai/l. Muse. Brit. p. 2. t. viii.
Hab. Hermite Island, Cape Horn, frequent on maritime rocks.
Slightly differing from British specimens in its wider, less evidently papillose leaves, which are more suddenly dilated near the middle. Bridel (Bryol. Univ.) cites with doubt, Dillenius (Muse. p. 507. t. 73. f. 40) as a synonym for this species, but the description of Dillenius distinctly mentions the nerved leaves which are characteristic of A. Rothii, to which species even the description of Linnaeus (as Bridel properly remarks) seems to refer. The secund foliage, generally ascribed to A. rupestris, is not a constant character, even in specimens gathered in the same locality.
