Thoughts on the Hymenoptera Trichogrammatidae etc. 71 



down through a common heritage. In the Classification of any group, 

 whether of genera or larger segregates, there is an underlying principle 

 that of common descent, which expresses itself in the comparative 

 ßtability of various characters. If a character is constant it takes high 

 taxonomic value, whatever the nature of the part may be. Its value 

 is distinctly dependent upon the number of species or genera it is uni- 

 form and constant for and also in its being correlated with other con- 

 stant characters. The theory of descent with modification also shews 

 why an aggregate of characters is of still greater importance in Classi- 

 fication, since it is clear that several constant characters give greater 

 evidence of common blood relationship. It is impossible for a character 

 to be constant in a group without its being evidence that the included 

 species are akin. Classification then is a matter of evidence of relation- 

 ship between species and their groupings. On this view we cannot con- 

 clude otherwise than that there are natural genera,^) 



The Classification of the Trichogrammatidae must be based upon 

 the principle stated and therefore I have taken constancy of a character 

 or of characters as an index of a natural or true grouping of the genera. 

 There is no character which is constantly correlated with others. Thus 

 all the members of the family so far known either have curved or 

 straight venation, but neither kind of venation is correlated with any 

 other character; all the members of the family have wings whose 

 discal ciliation is either normal and dense or eise arranged in regulär 

 lines and less dense and with neither kind of ciliation is there cor- 

 relation^); all members of the family bear antennae which are either 

 with or without a funicle, but neither division thus formed has correlated 

 characters; and all members of the family either do or do not bear 

 exserted ovipositors but neither those which do or those which do not 

 bear any other distinctive character excepting that most genera that 

 do, have the ciliation arranged in regulär lines. There is no other 

 distinctive character excepting that all genera with normal, dense 

 ciliation do not bear an oblique line of discal cilia from the stigmal 

 vein (but there are several unlike genera with regulär lines of ciliation 

 not possessing the oblique line of cilia, e. g. Tumidiclava, most species 

 oiAbhella, the species of Trichogrammatoidea, Laihromeris and Uscanella; 

 this cannot bi considered a correlation). Thus there is no other Operation 

 necessary in this case than a determination of the relative taxonomic 

 value of these four characters as expressed in their constancy. Let us 

 examine the last first. The presence of an exserted ovipositor is least 

 in importance of the characters named, since firstly, it is a character 

 of an adaptive nature as in Poropoea and Ophioneurus which deposit 



') How are we going to deny that genera are natural ? They are bom 

 in nature like variations and species are. There are natural orders as there 

 are natural varieties. To deny one is to deny the other, since to deny a 

 whole made up of similar parts, granted to be natural, is to deny nothing. 



^) One Australian genus is an exception, the ciliation irregulär and not 

 very dense; its ciliation is exceptional. 



1. H«n 



