DISTRIBUTION OF SPECIES 



nearest mainland leaves little doubt that the island dwellers have mi- 

 grated from the mainland. However, unless an island is very close to the 

 mainland, the difBculties of crossing the water barrier will keep out a large 

 portion of the potential migrants. Thus competition is less rigorous, and 

 the selection pressure lower than on the mainland. In the absence of the 

 check of normal interspecific competition, and often in the absence of 

 predators, the populations may reach prodigious numbers relative to the 

 space available. Yet the total population is small as compared to conti- 

 nental populations, and it is likely to be broken up by ecological factors 

 into much more restricted breeding colonies. In such a situation, genetic 

 drift may be stronger than selection, with the result that island forms are 

 likely to be less well adapted than their mainland relatives. Apparently 

 (but not actually) contradictory to this is the fact that closely related 

 island dwellers may become adapted to situations so different that only 

 remotely related organisms of the mainland could be compared to them. 

 An example of each situation will be discussed below. 



Genetic Drift on Oceanic Islands. A celebrated example of the former 

 type is that of the snails of the genus Partula on the islands of Tahiti and 

 Moorea, neighbors in the Society Islands, about 2400 miles south of 

 Hawaii. These are typical volcanic islands, characterized by a central vol- 

 cano from which deep valleys and narrow separating ridges radiate to the 

 sea. The snails feed on the plants of the valleys, and the intervening ridges 

 are almost impassable barriers to them. Several species of Partula are rep- 

 resented on the islands, some species being found in many of the valleys, 

 others in few or only a single valley. In each case, the inhabitants of every 

 valley comprise a distinct race, characterized by such things as size, direc- 

 tion of coiling, details of shape, and color. Only those species which are 

 represented in only one valley, such as P. tohiveana, are monotypic. The 

 variation of the several races of a species definitely is not clinal. Races of 

 neighboring valleys may be strongly divergent, while races at opposite 

 ends of an island may be quite similar. Different species in a single valley 

 may have identical ecological requirements, living on the same food 

 plants, and yet they may show no tendency toward parallel variation. All 

 attempts to interpret this situation in terms of different selective forces in 

 the different valleys have failed. It appears probable, then, that all of the 

 races of Partula are subject to substantially identical selective forces, and 

 the differences between them result from genetic drift. But genetic drift 

 on so grand a scale is possible only because the geographic features of 

 the islands enforce almost complete isolation upon all of the local breeding 

 populations. These snails have been studied intensively three times, at 

 wide intervals, in the past eighty years, and it appears that significant 

 changes in many of the races have occurred even in that short time. 



Rapid Selective DifFerentiation on Oceanic Islands. One of Darwin's 

 studies in the Galapagos Islands was made upon a family of finches, the 

 Geospizidae, which had become adapted to an amazing range of eco- 

 logical niches in the islands, so that their superficial differentiation was 

 much greater than usual within the confines of a single family. Another 

 family of birds, the Drepaniidae or Hawaiian honey-creepers, has under- 



S09 



