PHYSICAL GEOGRAPHY OF THE ANDES. 107 



of the species if compared to what is found in some 

 other mountain districts. If we were to find in South 

 America anything Hke the variety of species limited 

 to very small areas that is encountered in Southern 

 Spain, Greece, Asia Minor, and Southern Persia, where 

 on each mountain that we ascend we find several 

 well-marked local species, differing from those in 

 similar stations a few miles distant, the catalogue of 

 the Andean flora would have to be extended to three 

 or four times its actual length. 



Fully agreeing, as I do, with Engler in his general 

 conclusion that dry climates are more favourable than 

 moist ones to the development of new varieties, which 

 are the ancestors of future new species, I must remark 

 that in the Andes, so far as we know, the species with 

 very restricted area abound more in the upper zone, 

 where the soil is relatively moist, than in the drier 

 middle or lower zones. Nor does it appear that isola- 

 tion of the summits can be with reason invoked as an 

 explanation. The most marked feature in the range, 

 and one that geologists have perhaps not taken enough 

 to heart, is the extremely continuous character of the 

 crest of the range, especially on the western side, as 

 is evidenced by the fact that from Colombia to Southern 

 Chili there are so very few passes below the limit at 

 which snow frequently lies on the surface. For a 

 rational explanation of the facts as to the distribution 

 of mountain floras, we are forced to assume that the 

 various agencies which are in daily operation — birds 

 and land animals, winds, etc. — are competent to effect 

 the transference of the great majority of species from 

 one mountain to another not very far removed ; and 



