OF THE PERUVIAN ANDES. 



17 



JBicarpellafce. 



Oosniopolitnn genera 

 Amphigean genera , 

 American genera ..., 

 Antarctic genera . . . . 

 Andean genera 



Totals 



Chloris Andina. 



Rimac 



Valley. 



No. of 



No. of 



No. of 



1 



No. of 



1 



genera. 



species. 



genera. 



species. 



1 

 1 



19 



154 



10 



24 



1 



6 



1 



1 



8 



28 



10 



10 



2 



46 



2 



6 



18 



39 



4 



4 



48 



273 



27 



45 



Allowing for the numerous circumstances leading to discre- 

 pancies in the results of sucli a comparison, the figures given 

 above show a remarkable amount of agreement, and lead to tbe 

 inference that (when the special elements of the Chilian flora are 

 excluded from consideration) the small collection made by me is 

 fairly representative of the general character of the flora of the 

 Peruvian Andes. I have therefore thought it worth while to 

 consider what indications it may afford to elucidate the problems 

 of the past history of the South-American flora, availing myself 

 where possible of the fuller materials supplied by the ' Chloris 

 Andina.' 



The first fact which must be taken into account is the prepon- 

 derance in this flora of types that are very widely, almost uni- 

 versally, diffused throughout the world. It lias been already 

 pointed out that nearly nineteen twentieths of the species (193 

 out of 206) belong to 40 natural orders that are found in almost 

 all the mountain-regions of the world. Comparing the flora of 

 New Zealand, as one of the most distinct and isolated, I find that 

 only two of the above 40 orders, Ca^paridecs and AsclepiaJece, are 

 wanting in that region. I further note that (omitting Datisceco) 

 more than 60 per cent. (12G out of 206) of the species in my col- 

 lection belong to 62 cosmopolitan genera, many of which have a 

 range nearly as wide as the natural orders to which they belong. 

 These proportions would doubtless have been somewhat altered 

 if I had been able to collect more largely in the higher region. 

 I should have added largely to the number of generic types 



LINN, JOURN. BOTANY, VOL. XXII 



C 



