386 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. 57. 



project as originally conceived, which was to revise the North 

 American Pimplinae. 



It seems very probable that the Ichneumoninae as here treated is 

 composed of a number of phylogenetic branches, associated by more 

 or less similarity of hal)itus and superficial structure, and forming a 

 group which, because of its discordant elements, is so closely re- 

 lated to groups in other subfamilies as to make definition of it as a 

 unit extremely difficult if not impossible. The same is true, to 

 greater or less extent, of the other four subfamilies. Because of 

 this, it does not seem advisable to offer any other definition of the 

 limits of the subfamily Ichneumoninae than to say that this tribal 

 synopsis is based on the PimpHnae as limited by Ashmead and the 

 genera described since 1900. The student unfamiliar with these 

 insects will therefore necessarily have to refer to the unsatisfactory 

 definitions offered by Ashmead, Cresson, and others. 



In spite of the probable diversity of origin of the tribes, constant 

 diagnostic characters are very rare, and our keys to the tribes in- 

 clude the most available characters. Not all of these characters 

 taken singly are of tribal value; in fact such characters are very rare 

 in the Ichneumonidae. It is rather an assemblage of characters, 

 which taken together form the peculiar structure and habitus that 

 distinguish the members of one tribe from those of another. 



The remarks in regard to the subfamihes set forth above are 

 equally appUcable to the usual five tribes of the Ichneumoninae. 

 None of the existing keys to these tribes is usable by the beginner in 

 the taxonomic study of these insects unless he has access to an ex- 

 tensive named collection for comparison. This is due partly to the 

 attempt of the authors of the keys to place all the genera in the 

 five tribes originally proposed by Ashmead; partly to the use of 

 extreme or unisexual characters unsupported by characters applying 

 to the other sex, or of those variable within a genus, or the positive 

 statement of characters that do not apply to all species placed in a 

 given tribe; and partly to a too superficial study of the species, or, 

 when careful study has been made, a laissez-faire policy of not 

 attempting a revision of the keys. In other words, in order to run 

 a given species to its genus one must be able from the knowledge 

 gained only by long experience to tell at once the tribe to which the 

 species should be referred. The tribal keys, being unusable by the 

 beginner and unnecessary to the experienced, are useless, or worse 

 (so far as the beginner is concerned), misleading. 



We are convinced that the entire biology and the charactei-s dis- 

 played by all stages are of importance as indicating relationship or 

 divergence and that whatever of this sort of information is available 

 should be taken advantage of as a guide to classification. In pro- 



