No. 637] 



PARASITISM 



143 



pears that we have in them the most primitive parasites 

 in the order Hymenoptera. The hosts of the Oryssidae 

 consist partly, although probably not entirely, of Bu- 

 prestidas, which paleontology shows to be an ancient 

 family. Handlirsch ( '08) has even gone so far as to sug- 

 gest that the parasitic Hymenoptera may have been de- 

 rived from the Jurassic Pseudosiricidae which no longer 

 laid their eggs in wood, but in the eggs of beetles occur- 

 ring in the wood. This is entirely speculative and so I 

 think must be at the present time any suggestions as to 

 how the Oryssidae, or the Ichneumonidae, which Hand- 

 lirsch had in mind, became parasitic. That their larvae 

 first found and fed upon their hosts after hatching seems 

 much more probable. It must be said, however, that 

 predatory or carnivorous Chalastogastra are not known 

 among living forms, except certain adult sawflies which 

 fed in this w^ay (cf. Mrazek, '09). From this point on- 

 ward we have little trouble in tracing the probable origin 

 and relationships of the Ichneumonoid families as I have 

 attempted to do in a previous paper (Brues, '10). Thus 

 the Stephanidae are structurally primitive and strikingly 

 like the Oryssidae in the peculiarly horned head which 

 had been remarked on before the habits of the Oryssids 

 were knowoi. On account of the presence of a costal cell 

 in the wing, the polymorphic family Evaniidae is neces- 

 sarily also more primitive than the Ichneumonidae or 

 Braconidae, and, through one subfamily, the Foeninae, re- 

 semble the Stephanidae as has been already noted by 

 Bradley ('08). Some Braconidae, the Stephaniscinae, 

 Spathiinae and Hormiinae are much like Stephanids, so 

 much so that it is difficult to believe that they are not 

 directly derived from them. One other family, the Capi- 

 toniida^, recently segregated from the Braconidae, Tap- 

 pears to be very definitely related to the mor(^ general- 

 ized Evaniidae (Aulacin^). Omitting in this bi-icf ron- 

 sideration several less pertinent families, nud iu'.<MMiii>' 

 other recently segregated ones, we hav(> left (i!il> uw 

 Ichneumonidae, related possibly through the Alysiidii" to 



