486 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. 5s, 



Uiiitetl States National Museum, it is doubtful whether this could 

 have been accomplished at all. Information relative to the hosts 

 and cocoons is valviable, and sometimes aids greatly in tlie absolute 

 determination of a particular species. This information has espe- 

 cially helped the vs^riter to determine the extent to which various 

 structural characters could be relied upon for the separation of 

 species. 



In some sections of the genus color, especially color of the tegulae 

 and of the legs, can be safely used to distinguish species ; and color 

 of the dorsum of the abdomen and of the antennae are sometimes suf- 

 ficiently constant to be useful, but in other cases they are of no value. 

 The Avings supply but few good characters; however, the length of 

 the first abscissa of the radius as compared with that of the trans- 

 verse cubitus has been found valuable, as have also the color of the 

 wings, the length of the metacarpus as compared with that of the 

 stigma, and the direction of the nervellus. Leg characters, other than 

 color, that have been employed are the length of the inner spur of the 

 posterior tibiae as compared with that of the outer spur and with the 

 metatarsus, and the sculpture on the outer face of the posterior coxae. 

 The variation in the length of the female antennae has been found 

 useful in a few instances, and the length of the ovipositor, varying 

 from subexserted to longer than the abdomen, is often valuable in 

 separating individuals of this sex. But the most important dis- 

 tinguishing characters are found in the sculpture of the propodeum, 

 and in the shape and sculpture of the two basal abdominal tergites; 

 many considerable differences occur which are quite constant; and 

 yet even here the variation within species is often so great that other 

 supporting characters must be employed. In all cases it is very de- 

 sirable, when maldng determinations, to have before one a good 

 series of specimens that individual variations may be noted and fully 

 allowed for. 



Probably all species of Apanteles are parasitic exclusively upon 

 the larvae of Lepidoptera; at least there is no authentic record of 

 a species of this genus having been bred from a larva o^ another 

 order. Within themselves the species are by no means restricted 

 to a single host, but almost invariably the hosts are very closely 

 related, or at least there is a similarity of larval habit. Apanteles 

 congregatus Say, for example, has been reared from a large number 

 of hosts, none of these, however, being outside the family Sphingidae. 

 Apanteles xylinus Say, laeviceps Ashmead, grifflni Viereck, and 

 grenadensis Ashmead are cutworm parasites, but apparently make 

 no distinction between different species of cutworms; Apanteles 

 ornigis Weed appears to infest leaf-miners only, but is not restricted 

 to a particular species, attacking those forms which spend their 

 entire ]ar\ al life within a mine as well as those which are true 



