INTRODUCTION 11 



In determining the relative vahie to be assigned to different 

 characters for tlie pnrposes of classification, those characters 

 which are adaptive, i.e. liable to be affected by external agencies 

 through the means of natural selection, are generally very un- 

 trustworthy. Conversely, a character which seems of little 

 physiological importance, and not easily modified by ordinary 

 external influences, is commonly of especial value. Thus colour 

 and outline, the hairs of larvae, and the genital organs of 

 imagos are likely to be of slight importance in the definition of 

 groups ; whilst neuration occupies a high position, except when 

 directly influenced by an alteration in form of wing, which is 

 seldom the case. A combination of several charactei's which 

 are apparently not interdependent carries very great weight. 



It is sometimes said that genei'a (and families) are artificial 

 creations. If by this is intended that they consist of a certain 

 number of species having no other relation than the common 

 possession of certain characters, the statement is not true of 

 any sound system, and the systematist who makes it stands 

 self-condemned ; but if it is only taken to mean that the precise 

 limits of genera may often be differently conceived by different 

 workers, it is to that extent quite true. If two small allied 

 groups of species agree in all essential characters, and are 

 capable of definition as a whole which can be distinguished 

 from all other groups, yet are also constantly separable from 

 one another by a single and perhaps slight character, the 

 question whether they are to be reckoned as two genera or as 

 sub-groups of a single genus must always remain one which 

 may be answered differently by different workers. The same 

 is true of families ; and a consideration of the nature of a 

 genealogical tree and the diversity in the length of its branches 

 will show that it is unreasonable to expect all individual groups 

 of any particular denomination to be of the same actual value. 



Some writers have urged that secondary sexual characters 

 should never be used to define genera. No doubt there is a 

 convenience in only using such characters as are common to 

 both sexes, since the student who is trying to identify the 

 insect may only possess examples of one sex ; but there is 

 absolutely no scientific justification for the restriction, which 

 would make the classification of some groups of LejndojUera 

 quite impracticable ; when such structural characters are found 

 in any instance to assist the definition of natural genera, there 

 is no reason in nature why they should not be employed with 

 perfect freedom. 



