172 THE VARIATION OF ANIMALS IN NATURE 



value) on the ratio between the numbers of the variant and 

 the total number of individuals in the species. 



As we have said already, the actual basis of correlation is in 

 nearly all species unknown, but there are certain methods by 

 which important information may be obtained, indicating 

 that the correlation is often of the second type. 



(i) There may be considerable presumptive evidence that 

 the characters are physiologically independent of one another. 

 Thus in insects we should have no reason to suspect a direct 

 physiological relation between the arrangement of the wing- 

 nervures and the structure of the external genitalia, or, in 

 birds, between the shape of the beak and the colour of the 

 tail. How far the mere unlikelihood of a relation is significant 

 has to be decided in each individual case. The somewhat 

 anecdotal instances of correlation between apparently inde- 

 pendent parts which are cited by Darwin should be borne in 

 mind. 



(2) Some specific characters are unusually variable and 

 cannot, therefore, show a very high correlation with more 

 stable ones. Wherever low correlations are observed, there 

 is a likelihood that the basis is not physiological. More im- 

 portant evidence can be obtained in species in which in some 

 individuals a single specific character is replaced by one 

 normally distinctive of another species. The identity of such 

 aberrant individuals may be reasonably certain, since the other 

 members of its character complex are still associated together. 

 Further, these variant forms may be quite rare, so that the 

 correlation of the character in the species as a whole remains 

 high. Such cases strongly suggest that the character (and, by 

 inference, similar characters in allied species) is capable of 

 independent segregation. 1 



As an instance of this type of evidence we may mention the 

 Tortricid moth, Euxanthis straminea [cf. Waters, 1926, p. 159). 

 A form has occurred in S. Devon (and elsewhere) which, in 

 its large size and distinct dark wing markings, resembles the 

 allied species E. alternana. The aberrant specimens, however, 

 have typical genitalia, and the direction (though not the inten- 

 sity or dimensions) of the wing fascias is normal. Similar 



1 Nabours (1929, p. 33) has made the interesting observation that there are 

 differences in the linkage relations of similar patterns in different species of 

 grouse-locusts. 



