8 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.61. 



unwieldy; accordingly it seems desirable to recognize the artificial 

 division into three genera. Within each of these three groups many 

 species can be determined only with the greatest difficulty. Because 

 of the considerable variation of specific characters it is always very 

 desirable to have a series of several specimens. Odd collected speci- 

 mens, particularly if they be males, must frequently be placed with 

 some reservation ; this applies especially so Apanteles. 



The subfamily Micro gaste7'inae contains many useful species. 

 All that have been reared are parasitic on lepidopterous larvae, 

 and many are exceedingly important agents in the control of in- 

 jurious species. This is particularly true of numerous species of 

 Apanteles, as was pointed out in the publication to which reference 

 has been made above. 



The genus Apanteles has been omitted from this paper, except for 

 the description of four new species and brief synonymical notes. 

 For a treatment of this group the reader is referred to my recent 

 Revision of the North American Species of Apanteles.^ 



KEY OF THE GENERA OF MICROGASTERINAE. 



1. Anterior wings with the median and submedian cells of equal length ; both 



intercubital veins wanting, hence only one cubital cell present ; antennae 



with 24, or more, segments 2. 



Anterior wings with the submedian cell distinctly longer than the median ; 

 at least the first intercubitus present; antennae always with less than 

 24 segments 3. 



2. Antennae 2S-segmented ; mesoscutum normal, without a fovea on the disk. 



Oligoneurus Szepligeti. 

 Antennae 24-segmented ; mesoscutum with a rounded fovea on the disk. 



Mesocoelus Schulz, 



3. First intercubitus long, attaining the stigma ; radius not angled, obsolete ex- 



cept at extreme base 4. 



First intercubitus much shorter, not reaching the stigma ; radius with two 

 abscissae, the first making an almost right angle with the second, the 

 second usually only indicated by a line of closely placed setae 5. 



4. Antennae 14-.segmented ; stigma triangular, broad ; radius arising from middle 



of stigma ; second cubital cell not adjoining the first discoidal at base. 



Mirax Haliday. 



Antennae 20-segmented ; stigma elongate-oval ; radius arising far beyond 



the middle of stigma; second cubital cell adjoining the first discoidal 



at base Adelius Haliday. 



5. Antennae 21-segmented ; second abscissa of radius not entirely wanting ; 



propodeum regularly areolated Dirrhope Foerster. 



Antennae IS-segmented ; second abscissa of radius merely defined by a line 

 of setae; propodeum not regularly aerolated, at most with a median 

 areola G. 



6. Second cubital cell open behind, the second intercubitus entirely wanting. 



Apanteles Foerster. 



Second cubital cell not confluent with third, the second intercubitus present, 



though often more or less hyaline 7. 



» Pioc. i:. S. Xr.t. Mus., vol. 58, 1020, pp. 483-576. 



