RANGE BASE OR REFUGE AREAS 43 



they have never been connected with the adjoining 

 continent, and bases his behef chiefly on the absence 

 of all indigenous land mammalia and reptilia. I am, 

 however, inclined to believe that these islands have 

 experienced a continental period in remote ages, and 

 that they are but the isolated terminal portion of the 

 great Atlas range. The vast elevation of the ocean bed 

 required (some 5000 feet) to unite them now with conti- 

 nental Africa is not so much as Dr. Wallace himself 

 shows necessary to join the Philippines with continental 

 Asia (some 6000 feet), and but little more than what 

 must have occurred to connect certain land surfaces in 

 the eastern Mediterranean, of which we hav^e direct 

 palcEontological proof of union during Pleistocene time. 

 Further, it is a well-known fact that local marine de- 

 pressions of vast depths are by no means uncommon 

 features of regions subject to volcanic action. The 

 absence of terrestrial mammals and reptiles does not 

 appear such an insurmountable obstacle when we re- 

 member that the area during a past continental period 

 was a very isolated one, and that the adjoining parts 

 of Africa even during Pleistocene time were too arid, 

 desolate, and barren to support a very large mammalian 

 or reptilian fauna, and the probabilities are that it did 

 not contain any such animals at all in areas contingent 

 to the Canary group. I would suggest a submergence 

 (perhaps of a sudden, rather than of a gradual character) 

 some time during the Pleistocene Period, after such 

 desert-loving species as the Courser, Houbara Bustard, 

 and Sand Grouse (dominant desert forms) had gradually 

 spread westwards with the slow emigration of life over 

 the recently elevated Sahara. The avifauna of the 



