OF THE PERUVIAN ANDES. 13 



Audean flora, it will be best to exclude altogether the ferns, as 

 well as the species introduced by man. The 206 remaining 

 species include representatives of 49 natural orders. Of these, 

 five orders — FassiJlore(B,Bignoniac€ce, NyctaginecB^ PTiytolaccacecB^ 

 and Co77imelinacece, each appearing onlj by a single species Avhich 

 does not rise above the level of IO3OOO feet, are mainly inhabi- 

 tants of the tropical zones of the Old and New Worlds. Ilydro- 

 phyllacece^ represented by one species, and Folemoniacece^ by two 

 species, are orders whose original home appears to lie in the 

 north-western part of the American continent, and which are 

 sparingly represented in South America. Loasacece alone, of 

 which four species are enumerated in the following list, form a 

 group which may be regarded as specially Andine. With the 



^ liar exception of the monotypic genus Kisse7iia in east tropi- 

 cal Africa, the other genera of this family, including about 100 

 species, appear to have all originated in the mountain regions of 

 Central and Southern America, which I shall here designate as 

 the Andes, using that term in a broad sense. Setting aside the 

 somewhat doubtful representative of the small anomalous order 

 J)atis€€(S, we find that 193 species in the annexed list belong to 

 40 natural orders, which are all very widely distributed through- 

 out the world. 



Descending from natural orders to the categories of suborder 

 and tribe, we discover few indications of a special Andean flora. 

 The genus Balhisia represents the small distinct Andean tribe of 

 JVendtiecd belonging to Geraniacece^ and Malesherhia belongs to a 

 tribe of PassiJJoracece peculiar to the Central and Southern Andes. 

 The genus Calceolaria^ which by itself constitutes a tribe of 

 ScropliularinecE^ is represented at Chicla by five species ; but there 

 are very strong reasons for referring its origin to the Antarctic 

 region rather than to the Andes. To many a botanist the most 

 distinctive element in the Andean flora is afforded by the Muti- 

 siacecSy a tribe of the omnipresent CompositcB^ which is not, indeed, 

 confined, to the South American continent, but which there finds 

 its chief development, especially in the Chilian Andes. Of 57 

 known genera*, containing about 420 species, 42 genera and 

 fully 350 species are confined to South America. The rest are 

 scattered from South Africa to China, and only one American 

 genus {Trichocline) is represented elsewhere, i. e. in Australia. 



Throughout this paper I have adhered closely to the limits of genera as 

 defined in Bentham and Hookers 'Genera Plantarum/ and in the following 

 remarks I hare applied the same rule in comparing my results with those of 

 Weddell, 



