GEOGRAPHIC ISOLATION 87 



m. in depth) ; 54 and Lake Tanganyika, with an area of 35,000 sq. km. 

 and a depth of 1435 in. Recent studies of the fresh-water fauna of 

 Lake Ochrida in the Balkan Peninsula show that it is to be added 

 to this list, probably with some other Balkan fresh waters. 55 



In fauna! composition the Caspian Sea, with a surface of 439,000 

 sq. km. and a depth of 946 m., compares more closely with these bodies 

 of fresh water than to any of the partially isolated seas. 



Geographic isolation on land. — Islands are unquestionably the 

 most effectively isolated of land areas. They are comparable with 

 inland lakes and ponds in being surrounded completely by an effective 

 barrier; but they differ from these bodies of water in their relative 

 permanence. Islands therefore present an abundance of endemic species 

 and genera, and even some endemic families and orders. The amount 

 of endemism is inversely proportioned to the accessibility of the island 

 from other, more thickly populated areas, usually the nearest main- 

 land. The strength and regularity and direction of the wind are also 

 factors in populating islands. Endemism on islands is most frequent 

 in forms for which the difficulty of reaching the island is most extreme, 

 so that new increments of the parent form are unlikely to follow. 

 Endemic species are thus sometimes more numerous among mollusks 

 and reptiles than among insects, and more numerous in the latter 

 group than among bats and birds. The Antilles have a large fauna 

 of fresh-water snails in common, but their land snail faunae for the 

 most part differ from island to island; the eggs of the former are 

 easily distributed by aquatic birds whereas the latter are dependent 

 upon driftwood and storms. The converse relation may appear, as in 

 the Polynesian islands, where there are endemic species of birds while 

 the lizards are identical from island to island. Similarly in Melanesia 

 the fruit bats with powerful flight have formed numerous endemic 

 species in several archipelagoes. 56 



The duration of the isolation is naturally of great importance for 

 the amount of change undergone by the fauna of an island. Unfortu- 

 nately this very amount of differentiation is usually the only evidence 

 available for the estimation of the age of an island. In the very 

 recently repopulated Krakatoa (cf. p. 56) mo differentiation has taken 

 place. On ancient islands very completely isolated the genera are 

 frequently split up into species (cf. p. 520). 



The amount of endemism on an island is little influenced by the 

 mode of its origin, whether by the breaking up of a former continental 

 connection or by independent development as a volcanic or coral 

 island. The degree of isolation in either case may be the same. The 



