10 MR. J. BALL OX THE FLORA. 



As regards the proportion in which tlie natural families of 

 plants are represented in the Andean flora, the figures here given 

 agree pretty well with results derived irom other sources. The 

 number of Composite amounts to nearly a quarter of the whole 

 number of species, and this proportion seems to prevail through- 

 out the temperate regions of South America. But the propor- 

 tion in my list for the higher or alpine region is a full third of 

 the whole, and, as far as I can judge, this is not far from the truth 

 as rccfards the Andes of Peru and Bolivia. Weddell's ' Chloris 



\ 



Andina' is unfortunately incomplete, and it includes plants of the 

 upper region throughout the entire range from Columbia to Chili, 

 so that it does not allow of accurate comparison. In that work 

 478 species of Com^ositce are enumerated, and in the second 

 'olume, including all the remaining Garnopetalcd and nearly all 

 the Polypetalcje^ he reckons about 700 species. The remaining 

 phanerogams could scarcely amount to 400 species, so that the 

 proportion of Composites must approach to that above mentioned. 



Next to the Composifce the grasses play the most important 

 part in the Andean flora. In my li^t they amount to very nearly 

 one eighth of the whole collection, and in the alpine region to 

 one sixth. I am disposed to estimate the true proportion in this 

 part of the range at about one tenth. 



It is not to be expected that in so small a collection the number 

 of species belonging to the less predominant groups should cor- 

 respond nearly to the real proportions, yet on comparison with 

 other materials I have been surprised to find a considerable 

 amount of agreement. 



The third place in my list, as in Weddell's work, is taken by 

 the Scrophularinece^ which supply about five per cent, of the 

 entire number of phanerogams, and tlie fourth by Solanace<je, 

 while CrucifercB^ Caryop7iylJecB, and Leguminosee are each repre- 

 sented by 7 species, or one thirtieth of the whole number, probably 

 corresponding nearly to the true percentage. 



Conspicuous by its absence from my list is the family of Cype- 

 racecB. These are comparatively rare in the higher region of the 

 Andes, and apparently almost altogether absent from the dry 

 regions of Peru and Bolivia, ^rimulacece^ which are such con- 

 spicuous ornaments to the mountain floras of the old Avorld, are 

 everywhere rare in America, and their absence is not a matter for 

 surprise. 



The most remarkable feature in the local flora of Chicla is the 



