OF WESTERN SOUTH AMERICA. 159 



widely spread species of the hot and dry parts of the Mediterra- 

 nean region. The wide diffusion of nearly allied forms of this 

 genus in the desert regions of the Old "World, of Western IS'orth 

 America/aud of Chili, is an interesting fact of distribution ; and 

 it is perhaps still more remarkable that in Chili both forms seem 

 to be limited to a small zone and do not extend northward across 

 the Atacama desert or into Peru, Philippi found our species 

 no farther north than the neighbourhood of Caldera. Neither 

 of the Chilian forms is more different from Fagonia cretica than 

 are several described Oriental species which Bentham and Hooker 

 are disposed \o regard as varieties of the Linnean type. One 

 is tempted to speculate on the possibility of an early introduction 

 of the European species by the Spaniards and subsequent modi- 

 fications of the type. 



LoASA ciiiLEKsis = Huidobria chilensis, (7. Gay {Phil. 110). 

 This appears to be strictly confined to the Atacama region. Al- 

 though, as Bentham and Hooker have sliown, there are no suffi- 

 cient grounds for separating this generically from Loasa, it is a 

 very distinct species with a peculiar habit. The flowering 

 branches are often nearly glabrous, with very few of the glochi- 

 diate hairs characteristic of the genus. 



Gynoplettra RUGOSA=iMaleslierbia rugosa, C. Gay, Three 

 species, all allied to MalesJierlia himiliSy Don, and inhabiting the 

 arid region of Northern Chili, were described in the * Flora Chi- 

 lena,' and to these Philippi added three more found by him in the 

 Atacama desert. Further observation is required to ascertain tlie 

 constancy of the diagnostic characters adduced by these writers. 

 My specimen closely agrees with those of 3L rngosa received 

 from the Paris Museum, wliere 0. Gay's herbarium is deposited. 



Cruikshanksia nxMENonoTf, Hook, et Am. ? I cannot distin- 

 guish this from the original specimen preserved in Kew Her- 

 barium, but I am unable to say whether the imperfect specimen 

 given to me has aj^erennial root. G. tripartita^ Phil. Fl. Atac. No. 

 165, differs chiefly in being an annual. It was found by Philippi, 

 who does not enumerate G. hymenodon, at several places near 

 Copiapo; but I have not seen specimens of his plant. 



BACcnAHis co>'rEnTiFOLTA, Colla i?i Mem. Acad. Tier, xxxviii. 

 15 (1835) {PhiL 193)=B. ChW^uilla, JDG. Prod. v. 419 (1830) 

 B. margirata, DG. Prod. y. ^02 (1830) =B. parviflora, Pers. 



