EVIDENCE FROM DISTRIBUTION 141 



we can understand why this should have happened, 

 but, in a given case, it is not possible to do more 

 than conjecture the reasons. 



Another series of observations and deductions, 

 which depend for their cogency little or not at all 

 upon the data of geology and palaeontology, is that 

 afforded by the phenomena of the distribution of 

 animals and plants on islands. Darwin's attention 

 was strongly arrested by what he observed in the 

 Galapagos, and A. R. Wallace, who, it will be remem- 

 bered, independently reached the same conclusions 

 as Darwin, was led to his results largely by his studies 

 in the East Indian islands. To the end of his life 

 Wallace was fascinated by this topic and his two de- 

 lightful books, "The Malay Archipelago" and "Is- 

 land Life,' testify to his unfailing interest in its 

 problems. The problems of insular life are much 

 simpler and more comprehensible than those of 

 continental distribution; they are like experiments 

 performed upon a small scale and with relatively 

 few factors involved. 



From the point of view of animal and plant dis- 

 tribution there are two clearly distinguished classes 

 of islands: (1) Continental and (2) Oceanic. Con- 

 tinental islands are, as their name implies, detached 

 fragments of a continent, with which they were at 

 one time joined; their geological structure is that of 

 the continent and usually they are not far from the 

 parent continent and are surrounded by shoal water. 

 It should be explained that the line between shallow 



