OF WASHINGTON. 215 



with but few exceptions, are primary parasites on destructive 

 Lepidopterous larvae. 



This group has always been considered the most difficult one 

 to study in this family, and up to the present time remains in the 

 greatest confusion, as no oi>e has been able to seize hold of 

 salient characters that will readily separate the species. 



This, I think, can now be done, without much difficulty, from 

 the use of characters pointed out to-night. 



In my remarks under the thorax I have already called atten 

 tion to the great differences observable in the metathoracic 

 characters of this group, and these differences taken in connection 

 w r ith structural peculiarities of the abdomen, now to be pointed 

 out, will enable any one to readily separate these insects into 

 sections, subsections or divisions, and then into species. 



In figs. 18, 19, 20, 21, and 22 I show the structure of only 

 five different types of abdomen observable in this group, although 

 many others exist. 



On the first or basal segment of all the species in this group, 

 as well as sometimes in some species of other groups, is a raised 

 chitinous shield or plate of various shapes or forms. 



This plate I find to be exceedingly valuable in classification, 

 and I have made extensive use of it in my forthcoming mono 

 graph of the North American Braconidas. 



This plate, you will notice in fig. 18, is about three times as 

 long as wide, the sides being parallel ; in fig. 19 it is trapezoidal, 

 scarcely as long as wide, or a little wider at apex than at base ; 

 in fig. 20 quite a different shape is seen ; here it is much longer 

 than wide, or long trapezoidal, with the hind angles rounded ; 

 in fig. 21 we have a quadrate-shaped plate, covering almost 

 entirely the upper surface ; while in fig. 22 still another quite 

 different shaped plate is seen ; here we have a narrow, sharply 

 pointed, wedge-shaped or lanceolate-shaped plate. 



These are only a few of the many forms occurring in this 

 group. 



In this connection I would also call attention to another char 

 acter of primary importance, which should be more extensively 

 used in classification ; that is, the relative length of the tergites 

 or dorsal segments. 



