16 BULLETIN 1363, U. S. DEPARTMENT OF AGRICULTURE 



Harris (5, p. 2 IS). The writers have never succeeded in rearing this 

 parasite, but they have obtained records of two Hymenoptera, Am- 

 blyteles caliginosus Cress, and Hoplismenus morulus Say. 



The species Aglais milberti, Vanessa atalanta, Vanessa huntera, 

 and Polygonia convma are nearly always found just a few feet above 

 the ground, feeding on nettle and everlasting. The two former 

 usually feed in colonies, the two latter are apparently not so gregari- 

 ous. All are favorable hosts. 



In its struggle with the native tachinids, there is little fear that 

 Compsilura may not be fully capable of taking care of itself; but 

 it is less certain that it can successfully compete with certain 

 Hymenoptera. In the collections of Aglais milberti (1917, 1918) 

 there are but two parasites of importance, the hymenopteron 

 Apanteles atalantae Pack, and Compsilura. A. atalantae is gre- 



farious and is reared more frequently than the tachinid. The latter, 

 owever, breeds freely upon this host, and it is a frequent occur- 

 rence to rear five or six from a single individual. A few rearing 

 records showing the best examples of superparasitism are as follows : 



From a collection of 143 larvae, 34 were killed by Apanteles ata- 

 lantae, and from the remaining larvae (109) there were reared 310 

 Compsilura. No adults of Aglais milberti were secured. From a 

 collection of 101 larvae, 74 were killed by Apanteles atalantae and 

 from the remaining larvae (27) there were reared 41 Compsilura. 

 A single adult Aglais milberti issued. From 19 larvae, there were 

 reared 53 Compsilura. No host adults issued. From 26 larvae 

 there were reared 40 Compsilura. No host adults issued. 



Of the two parasites, Compsilura and Apanteles atalantae, the 

 latter has much the advantage, because of its ability to attack the 

 host in an earlier stage. It is probable that, if any parasite which 

 is capable of attacking the early-stage caterpillars should compete 

 with Compsilura, the latter species would be the loser. This is the 

 case as regards A. atalantae, for here there is a considerable number 

 of progeny already well advanced in development before the host is 

 subject to the attack of Compsilura. It is assumed that if the host 

 already parasitized by A. atalantae was in turn attacked by Comp- 

 silura there would be but little chance for its development, since the 

 parasites would crowd it out. This opinion is strengthened by the 

 absence of multiple parasitism in this species. A. atalantae attacks 

 the larva in the first and second stages only. This habit, as pre- 

 viously noted, has its advantages ; and yet, because of it, those indi- 

 viduals which have reached the third stage and have escaped the 

 attack of that parasite are free to develop without apparent check. 11 

 It is from the third stage on that Compsilura proves its worth ; and, 

 although it probably wastes many of its progeny in futile attempts 

 at parasitism on some already parasitized host, it nevertheless ac- 

 complishes its purpose and must be credited with a considerable 

 amount of parasitism. 



Vanessa atalanta L., a species closely allied to Aglais milberti 

 and similar in habits and behavior, is at times heavily parasitized 

 by Compsilura. For example, a collection of 25 last-stage larvae 



n This fact, and also taking into consideration the rearing of a few specimens of 

 (Exorista) Zenillia futilis O. S. and Winttiemia quadripustulata Fab., would perhaps 

 indicate some interference from Compsilura. 



