430 Mr. E. Meyrick on the classification 



natural separation ; on the contrary, I cannot but think 

 that no natural line of separation between Eastern 

 Siberia on the one hand, and Japan and North China on 

 the other, is capable of being drawn. Probably, how- 

 ever, on accurate investigation, we should find that there 

 is no natural line anywhere. 



It will be well to mention here some of the general 

 rules of classification. No genus, family, or higher 

 group, is tenable unless distinctly separable from all 

 others by points of structure, which, whether singly or 

 in conjunction, are capable of accurate definition. If a 

 systematist is not able to define by a clear and not 

 simply comparative character the distinction between 

 two genera, he is bound to merge them together ; thus, 

 to say that in one the cell is short and in another long, 

 is no sufficient definition ; to say that in one the cell is 

 less than one-third of the wing in length, and in another 

 more than one-third, is sufficient, if found constant and 

 clearly perceptible, but in practice it would probably be 

 a very bad character, as probably some species would be 

 transitional. Even where transitional forms are not 

 known, it will always be necessary to use judgment 

 as to whether the distinction employed is of such a 

 character as to be likely to hold good in the event of the 

 discovery of additional species. But even where there 

 is a good and definable point of distinction, it does not 

 follow that the genera are to be maintained ; where 

 genera are small and numerous, it becomes intrinsically 

 undesirable to multiply them, and in such a case, if two 

 small genera agree in nearly all structural characters, 

 resemble one another superficially, are apparently closely 

 connected genealogically, and finally are capable of 

 accurate definition and distinction as a single whole, 

 then they ought in general to be united. Many structural 

 characters are variable, either in difi'erent specimens of 

 the same species, or sometimes in a transitional series 

 of closely allied species. I hope shortly to give a paper 

 on the classification of the European Geometrina, and 

 shall then give some remarkable and, I believe, unpre- 

 cedented statistics of the variation of structural cha- 

 racters, but many instances will be found in the 

 following genera. The same point of structure will often 

 be found available as a good and reliable distinguishing 

 character in one instance, and not in another ; this can 



