C. SKOTTSBERG 



Masafucra 13 (28.3 %), on Masatierra and Masafuera 11 (23.9 %), and on all three 

 islands 3 (6.5 %). However, several of the species restricted to Masatierra are, 

 perhaps, not truly indigenous, whereas the indigenous character of the species 

 only recorded from Masafuera cannot be doubted; 11 of them are mountain plants. 

 With regard to the former, some were commented upon by Joiiow. In his table 

 "Continjente B, Especies autoctonas, pero no endemicas", pp. 221-222, he men- 

 tions several species regarded as doubtful natives: 3 species of GiiapJialiuvi, Mimu- 

 liis, Soliviuin fnrcatnin ( = robinsonianum), Tetvagonia cxpansa, Spcrgularia rubra 

 (= confertiflora), Parietaria hiiinifusa\ Trisetii])i cJiroiiosiacliyiivi and Da)itlionia 

 collina. With the exception of GiiapJialium spp. I have listed them as native. 

 There seems to be little reason to regard Tetragoiiia, a wide-spread thalassocho- 

 rous plant, as anthropochorous, and the Spergularia is known only from Juan 

 Fernandez and Desventuradas Is. Dauthonia gave me the impression of being 

 just as autochthonous as Piptochactiuiii and the two species of Siipa, and I find no 

 good reason to exclude either Mii/ai/ns, Parietaria or Trisetuiii (this never seen 

 by us). With regard to SolanuDi "'furcatuiir (not furcatum of Dunal) I share 

 JOHOW's doubts. It was described by BlTTER as 5. rohinsonianuni, a Morella micro- 

 species of ^. iiigruni L. coll. and found on all three islands. This assemblage 

 as represented on the mainland has never been seriously studied. Possibly ^. robin- 

 sojiiaJiiDH is a case of the same kind as the dandelion discovered on the islands 

 in 191 7 and described as Taraxacii))i fernandezianiuii Dahlst., a microspecies of 

 the Eurasiatic Vulgaria. There cannot be the slightest doubt that it is an alien 

 introduced from Chile, where these weeds have not been studied. From Joiiow's 

 list I have excluded Bahia, Amblyopappus, Erythraea, Mojwcosjiiia and Phalaris. 

 Further, some species not recorded from Juan Fernandez by JOliOW or earlier 

 authors, but for the time being listed as native by me, are under strong suspi- 

 cion: y uncus capillaceus and planifolius, Ccntella and Pledyotis, perhaps also Paro- 

 nychia. If, in the future, we shall be able to purify the list, the percentage of 

 endemic species will rise to 72 % or even more. 



The genera richest in species are Erigcron with 6, Juncus, Robinsonia, Unci- 

 nia and IVahlenbergia with 5, and Dendroscris and Peperomia with 4 species each; 

 eight genera have 3, 15 two, and 59 only one species. The average number of 

 species to a genus is 1.65. 



The systematical position of the endemites and the distribution of the genera and 



species also found elsewhere. 



Gramineae. 



Stipa L. About 250 sp. (Bews), widespread both hemisph., subtrop.-temp. 



neesiana Trin. et Rupr. Mex. to Boliv. and Chile, Braz., Argent., Urug. Poly- 

 morphous (2^^.771). 



lacvissima (Phil.) Speg. Peru, Chile, Argent. With a number of S. Amer. sp. 

 referred to a separate genus [^NasselUi). 



