1916] Williams — Neiv Species of Thripoclenus 59 



normally a rather long period of sucking the wound, followed oc- 

 casionally by a second oviposition. This last, however, was 

 seldom allowed to occur, as I desired to ol)tain as many parasitized 

 thrips as possible, and usually replaced the infested individuals with 

 fresh ones. There was no particular point selected for attack, 

 though the abdomen was more often pierced than the thorax. 

 In the length of time required for oviposition the species differs 

 markedly from T. russelli, which is described as taking less than a 

 minute for the process. The subsequent tasting of the wound 

 appears also to be omitted by the latter species. 



The number of larvse which a parasite would attack was not 

 definitely ascertained. After attacking three or four the female 

 Thripoctenus usually died or made its escape. Those that had 

 already oviposited could be counted upon for a repetition of the 

 process more surely than those which had not so far attempted it. 



All stages of the thrips except the adult are attacked, the new- 

 hatched larvse with an alacrity which appears to be ill-judged, as 

 the larvte of the parasite can hardly develop in them properly, 

 the pupae with a reluctance which may be well-founded. At 

 least one pupa transformed after a Thripoctenus had oviposited 

 in it. 



Development oj the Larvoe: The lengths of the egg and larval 

 stages of the parasite were not determined. It seemed impossible 

 to keep the larval thrips alive after removal from the galls, and 

 moreover, individuals oviposited in under laboratory conditions 

 often proved to have been previously infested. The only method 

 of avoiding this source of error was to raise the hosts from the egg, 

 and this could not be managed. Only occasionally were parasite 

 larvae obtained by dissection. These were extremely delicate in 

 the early stages. One well-grown specimen, dissected from a 

 pupal thrips, resembled the mature forms, being quite smooth, 

 very faintly segmented, and without any pigmentation in the head. 



Parasitization was not obvious until it had reached an advanced 

 stage. At this time the host larva grew mottled and waxy in 

 color, the body was much swollen, and the insect was perfectly 

 torpid, if not dead. In a day or two the outlines of the parasites 

 showed beneath its skin. Owing to the large size of the host, suc- 

 cessful multiple parasitism is possible, and indeed appears to be the 

 rule. The plate shows a thrips containing three larvae, and judg- 



