86 THE VARIATION OF ANIMALS IN NATURE 



animals by Riley (1924, p. 77). We hardly believe it feasible 

 to test that hypothesis with reference to the area occupied 

 by related subspecies. Willis's theory seems to have a partial 

 validity ; but, as Robson (1928, p. 114) has suggested, we are 

 not justified in dealing with it as of prime importance in 

 explaining differences in distribution. 



Fig. 7. — Distribution of Primary Varieties of Partula otaheitana on Tahiti. 



(Text-fig. 7 in Crampton, 191 6.) 



6. All taxonomists and probably very many other students 

 know that closely allied species are frequently united by 

 ' intermediates ' or, to put it in another way, that they have 

 different means but overlapping ranges of variation in some 

 characters. Other closely allied forms appear to be sharply 

 distinguished in all the characters investigated, though, of 

 course, the analysis is rarely pushed far enough to enable us 

 to say if such distinctions are found in every character. 



That all species have a certain, if sometimes very limited, 

 range of ' continuous ' variation is too well known to require 

 documentation. The notion of ' continuous ' variation is 



