THE PHANEROGAMS OF THE JUAN FERNANDEZ ISLANDS III 



plana, plerumque 5—8 mm lata, margine dense ac longissime albovillosa, apice 

 acuta subpungentia. Inflorescentia erecta, ddulta subnutans, ovoidea — cylindrica, 

 sat compacta, foliis 2 — 3 sat longis et latis instructa, interdum spicula infima 

 remota et longius pedunculata; spiculae compositae ovoideo-cylindricae vel 

 conicae, bracteis bracteolisque pallidis subhyalinis longe ciliatis. Flores ad 

 3 mm longi, tepalis integris, externis ovato-lanceolatis, ± convolutis, apice mani- 

 feste aristato-acuminatis, 2,8—3 mm longis, castaneis margine pallido, internis 

 2,5—2,6 mm longis acute acuminatis. Stamina 3 perigonio multo breviora, 

 antherae oblongae 0,4 — 0,5 mm, filamentis dtiplo breviores. Stylus brevis, 0,3 mm 

 longus, stigmata longa. Capsula nigro-castanea, trigono-sphaeroidea, apiculata, 

 perigonio multo brevior. Semina minuta oblique fusiformia, ferruginea, basi et 

 apice luteola, 0,8 — 1 mm longa. 



Masafuera: In the alpine region, rare, — The high ridges north of the 

 Casas valley, from 1000 m upwards; more common at Las Torres, 1350 — 1370 m 

 (fr. 14 /2 — 20 / 2 J 7> no - 4° 8 ); west wall of the island; C. del Barril', 1290 m. 



L. racemosa was established on specimens from the high volcanic peaks of 

 Mexico, collected by Humboldt and by ScklEDE, see Kunth Erium. PI. III.313. 

 L. vulcanica Liebm., of which I have seen authentic material (Herb; Upsal.) 

 seems to be the same and was also referred to L. racemosa by Buchenau. 

 These plants agree very well with the description, so I conclude that the 

 Mexican species is the true racemosa. KUNTH did not know this from Peru 

 or Chile, nor was it included in Gay's Bot. VI. BUCHENAU, in Pflanzenreich, 

 Juncac. 75 gave as area of distribution Mexico — Chile, and there are specimens 

 from Fuegia in Herb. Upsal. under the name of L. racemosa, determined by him. 



The position assigned to the insular form is entirely provisional. I am 

 Sure it has little or nothing to do with the Mexican plant, which has very narrow 

 leaves, a slender, almost filiform culm, long ciliated tepals, very short stamens 

 (less than half the length of the perigone) and a light brown capsule with very 

 thin, papyraceous walls. If we had only the Mexican plant to consider, the 

 form from Masafuera would never have been classed as a subspecies of L. race- 

 mosa. But according to Buchenau this is also a Chilean plant, occurring in 

 several forms down to" Fuegia, all, however, unlike the Mexican one in habit. 

 Many authors and collectors have followed Buchenau in bringing specimens 

 from the southern Andes to L. racemosa. In 1908 I collected such forms in 

 West Patagonia and in Skyring Water. L. *insularis differs from them as 

 well as from all other forms from the continent, that I have seen, in the broader 

 leaves, the more compact panicle, the longer tepals etc. On the other hand, 

 the forms of Chilean racemosa differ from each other in several respects, such 

 as the length of the filaments and size of the anthers, which are twice as large 

 in the form from West Patagonia as in the Skyring plant. Again, one plant 

 collected in Fuegia by DUSEN (no. 584) and determined by Buchenau has 

 the tepals hardly longer than the capsule, thus recalling L. chilensis. I have 

 not seen authentic material of the latter species. It differs from L. racemosa 

 in possessing 6 stamens, larger anthers (as long as the filaments), and larger 

 seeds (1 — 2 mm). 



It is probable that the Chilean L. racemosa differs from the Mexican; it 

 is evident that it includes more than one subspecies or species. When bringing 



