THYSANOPTERA AND ORTHOPTERA. 189 



Aeolothrips fasciatus, and probably Trichothrips heachi and Ctenothrips 

 hridwelli, fly often; Anaphothrips striatus, Chirothrips manicatus and 

 Scirtothrips rnthveni flew occasionally; Anihothrips. verhasci and A. 

 niger rarely; and Aptinothrips rufus, Allothrips megatephalus, Neo- 

 thrips corticis and Trichothrips hrevicruralis never. 



The first two species were the best subjects for observations on 

 flight. Large specimens of Thrips physopus found on dandelion {Tarax- 

 acum officinale) w'ere especially favorable if placed on white paper. The 

 method of flight is therefore described wholly from Thrips physopus 

 and Euthrips tritici. The individual did not usually fly from a steeply 

 inclined plane. It crawled up to the top of some eminence before at- 

 tempting to spread its wings. If, as soon as it had reached such an 

 eminence, its support were turned so as to leave the thrips at the bot- 

 tom instead of at the top, the insect usually turned and crawled up 

 again. The turning could often be repeated many times, until finally 

 the thrips apparently lost its hold and dropped. 

 . When the thrips had reached the point from which it was to take its 

 flight, it began to curve up its abdomen at frequent intervals (0.7 sec. 

 to 2.0 sec. were observed). The purpose of this was apparently to comb 

 out the long fringes which border the hind margins of both wings, and 

 which in the resting position of the w^ngs lie nearly parallel to the 

 longitudinal axis of the insect. There are many spines on the sides of 

 the abdomen which doubtless serve as teeth to the comb. The abdo- 

 men was usually curved up several times in succession, in some cases 

 eight or ten times. After this, I have seen a large female rest quietly 

 with her abdomen raised and wings spread for two seconds or longer 

 without any attempt at flight. Very often after combing out the hind 

 fringes with the abdomen several times, the wings were laid back in 

 resting- position, and the hind legs thrown over them. There followed 

 a series of movements somewhat resembhng those of a fly in cleaning 

 its AAings, during which the abdomen was slightly curved with its con- 

 vex side upward. Apparently only the hind tibiae covered the wings; 

 these tibiae are usually furnished with prominent spines. Further- 

 more it seemed that they did not extend to the posterior (inner when 

 at rest) edge of the wings, so that the tibiae combed out only the costal 

 fringes. 



Meanwhile the legs were placed in a rather definite position, the mid- 

 dle legs nearly at right angles to the long axis of the body, the other two 

 pairs very obliquely forward and backward respectively. Inasmuch 

 as the direction of flight was usually more nearly perpendicular than 

 parallel to the supporting surface, I judge that the middle legs are most 

 effective as aids in starting flight. This seems the more probable, since 

 in manv cases the hind legs had scarcelv finished combing the costal 



