184 Lelayid O. Howard 



from each other by certain structural characters, these seem 

 unimportant compared with those which we have been using, 

 and for the present, at least, these parasites must remain con- 

 generic. 



The single species reared from a beetle ^-g^ forms an isolated 

 type in the group as does also the single species reared from 

 an ichneumon cocoon. 



That reared from Ceratina is a single specimen lacking an- 

 tennse and these organs furnish the principal characters of the 

 European K. varicornis which we should expect it to resemble 

 from the fact that the latter was reared from Eumenes. The 

 other structural characters given are not especially distinctive, 

 but it is worthy of note that they agree with those of our 

 Ceratina parasite. 



We have then, in summing up, fourteen distinct types of 

 the genus Encyrtus to the discovery and exact definition of 

 which we have been led by a knowledge of the host-relations 

 of the species. Upon thirteen of these types new genera will 

 be founded, leaving to the fourteenth the old generic name. 

 Those parasitic upon Aphididse, Cynipidse, lepidopterous 

 larvae, coleopterous eggs, ichneumonid cocoon and bee larva 

 form each a distinct type. Those parasitic upon Coccidae 

 form three, two of which are well differentiated biologically 

 by the character of the host-insects within the family. Those 

 parasitic upon Psyllidae and Diptera form three and two re- 

 spectively, one of which is possessed by both in common, the 

 gall-making habit of the host producing the similarity in the 

 parasite, as is common in other parasitic groups. And these 

 parasitic upon heteropterous and lepidopterous eggs form a 

 single type, as is also occasionally the case with other parasitic 

 groups. 



This little paper then tends to show : (i) Another exempli- 

 fication of the axiom that structure is dependent upon habit ; 

 (2) that while the true classification depends entirely upon 

 structural detail, we may gain ideas as to the relative value of 

 characters by a knowledge of vital habits ; and (3) that as 

 soon as sufficient records accumulate it will be important to 

 examine the classificatory bearings of the group-habits, par- 



