REPORT ON THE BOTANY OF JUAN FERNANDEZ AND MASAFUERA. 63 



Trisetum (Kceleria) laxum, Philippi. 



Trisetum (Kceleria) laxum, Philippi, Descr. Nucvas Plantas, 1873, p. 92. 



Juan Fernandez. Philippi. 



There is no authentically named specimen of this plant in the collections examined, 

 and none that agrees with the description. 



Trisetum variabile, Desv. 



Trisetum variabile, Desv. in Gay FL Chil., vi. p. 351 ; Philippi in Bot. Zeit., 1856, p. 630. 

 Trisetum malacophyllum, Steud., Graminea?, p. 229. 



Juan Fernandez. Philippi; Moseley. 



Founded by Desvaux on Chilian specimens, with the remark that it is sufficiently near 

 Trisetum subspicatum — "Esta planta es bastante vecina del Trisetum subspicatum." In 

 the Kew Herbarium there is a specimen from Juan Fernandez collected, or at least com- 

 municated by Reed, and named Trisetum chromostachyum, Desv. 1 var. vaginatum. 



Steudel cites Bertero's No. 998 for his Trisetum malacophyllum, and Desvaux the 

 same number for his Trisetum variabile, so it may be assumed that they represent the same 

 plant. Indeed, there is little doubt that these and several other proposed South American 

 species belong to the generally diffused Trisetum subspicatum, Beauv. 



Pantathera fernandeziana, Philippi. (Flate LXI.) 



Pantathera fernandeziana, Philippi in Bot. Zeit., 1856, p. 649 ; Benth. et Hook., Gen. PI., iii. p. 1200. 



Juan Fernandez.- — In woods of the higher mountains — Bertero; Douglas; Reed; 

 Philippi; Moseley. Masafuera. — Germain. 



This remarkable grass is restricted to the islands, where, however, it would appear to 

 abound, for it is in nearly all the collections ; yet we find no description of it earlier than 

 Philippi's in 1856. But as Bertero collected it the probabilities of its being somewhere 

 described are great, especially as it is a large grass of distinct aspect. The specimens 

 vary in length from about fifteen inches with the root, to two feet and a half without the 

 root, and judging from the latter, the culms sometimes reach as much as three feet in 

 length ; but they are slender, and probably more or less procumbent. Among the numerous 

 specimens in the Kew Herbarium, it was difficult to find perfect flowers and grain, as they 

 are eaten by some insect, though fortunately those collected by Moseley furnish both flowers 

 and mature grain for the accompanying plate. Philippi mentions that his specimens 

 were in the same condition, and he was unable to describe either stamens or pistil. The 

 outer glumes have been described as three-nerved, but they are really five-nerved, the two 

 lateral nerves on each side being confluent from the base upwards to the middle of the 

 olume or even higher ; and the very coriaceous flowering glume is also five-nerved, though 

 inconspicuously. The spikelets are much laterally compressed, and the glumes consequently- 



