6 BULLETIN 38, UNITED STATES NATIONAL MUSEUM. 



nocturnal in habit, hiding under bark and in crevices during the day; 

 others are largely diurnal, flying in the bright sunshine, visiting flowers, 

 and not so usually taken at night. 



Tiie generic term Agrotis therefore conveys no distinct impression, 

 and it could only result in a huge assemblage of species through which 

 no path was visible and in which identification to any but the specialist 

 or to one with a large collection was all but impossible. Vague efforts 

 have been made from time to time to divide the genus and there exists 

 an abundance of generic names and types, but the true characters were 

 not recognized, species were erroneously associated, and the universal 

 result was that sooner or later the proposed terms went into the syn- 

 onymy and but aided to increase the already existent confusion. I have 

 made an eflbrt to use as far as possible existing generic terms in my 

 division of the genus; but I have not had exotic material sufficient to 

 give me an opportunity to study generic types, and the descriptions 

 •were in most cases useless, since they did not give the structural char- 

 acteristics upon which my scheme of division was based. I have there- 

 fore felt free to propose new names for the groups into which the spe- 

 cies have been formed, quite satisfied to have the names relegated into 

 the patient synonymy when the study of generic types proves their 

 identity with others already proposed. 



I had at one time the strong conviction that genera were natural as- 

 semblages, ca])able of strict limitation and definite in extent. The study 

 of very large material since that time has convinced me that my first 

 im^jression was erroneous, that genera as such are mere artificial divis- 

 ions of convenience, useful for the purpose of identification and for the 

 expression of relationship, and that they were useful for that purpose 

 just in proportion as they exi^ressed clear and definite associations of 

 characters. Agrotis is vague since it is definite on two points merely. 

 Carneades is definite, even though species are numerous, for it conveys 

 the association of a large series of characters in addition to that dis- 

 tinctive to it. The limits to which generic division may be carried will 

 depend entirely upon the tact and individual propensities of an author 

 and in large part also upon the tendency of the time at which he is 

 writing. After a period of wild generic creation there comes usually a 

 reaction in which lumping is carried to an extreme, and this is true as 

 to species as well as genera. 



For some time Mr. Grote was the only systematic worker in the North 

 American Noctuida?, and he steadily increased the number of generic 

 divisions until he was in advance of the allowable at that state of our 

 knowledge. The reaction was perhaps typified when, after becoming 

 thoroughly imbued with the critical ideas of my Brooklyn entomolog- 

 ical friends and teachers, I published my synopsis of noctuid genera. I 

 there went too far in the opposite extreme, but have no apology to 

 make for it. It acted as a check from whicli more careful studies have 

 allowed a steady, and I hope healthy and permanent advance. 



