REVISION OF STREPSIPTERA PIERCE. 63 



pupa on the second to the fourth, fifth, or sixth segments. These 

 invaginations deepen until they become long, funnel-like tubes, and 

 finally the inner end breaks, and through the canals thus formed the 

 triungulinids find exit into the brood canal which opens between 

 the head and thorax on the cephalothorax (Nassonow, 1892 e, pi. 

 2, figs. 18, 19, 20, 21; Nassonow, 1892 c, pi. 1, figs. 9, 11, 12). The 

 number of these median canals seems to be variable. In JCenos 

 vesparum (Nassonow) Nassonow found four only (Nassonow, 1892 c, 

 nl. I, figs. 2, 9). In Stylops melittse (Nassonow) he found five 

 (Nassonow, 1893 a, pi. 1, fig. 2; pi. 2, figs. 1, 6, 9). Muir found only 

 three in Elenchoides perkinsi Pierce (Muir, 1906). 



EXTERNAL STRUCTURE. 

 MALE. 



The integument of the male Strepsiptera is of a very thin chitin, in 

 many parts transparent, and very soft in all parts except the mandi- 

 bles and cedeagus. 



The head is transverse, with the eyes more or less stalked. The 

 eyes are composed of numerous hexagonal or circular ommatidia 

 separated from each other by densely ciiiate walls. Such a condition 

 occurs also in JEolothrips (Thysanoptera) and in Triclidbius (Pupi- 

 para) and in common parlance might be known as a "raspberry eye." 

 The eye partitions extend inward and separate the ommatidia as far 

 as the brain (Nassonow, 1892 c, pi. 1, fig. 16). Ocelli are absent (see 

 pi. l,fig. 9). 



The antenna? arise in front of the vertex between the eyes. They 

 are invariably flabellate in the third joint, and may or may not have 

 the succeeding joints produced. The entire structure of the an- 

 tennae is very sensitive and may be described as follows: The 

 surface is divided off into somewhat circular areas by stout, curved 

 bristles, with broad bases, which fence into each area a subcylindrical 

 tent of very transparent membrane surmounting a circular opening 

 into the interior of the antenna 1 . These tents are evidently highly 

 sensitive organs for communicating the necessary impressions to the 

 transient male. In Xenidae the areas are much denser and more 

 closely guarded than in the Elenchidse. In the Mengeidas the tents 

 are not as delicate, resemble tubercles, and bear the spines instead of 

 being circled by them. (See fig. 3 on p. 69, nos. 2, 4, 6, 7, 8, 10, 12, 

 14, 16.) 



The mouth parts are still very difficult of elucidation. A view of 

 the face shows a triangular or circular opening bounded by the nar- 

 row gena? and the vertex but open behind. At the base of this pit 

 is a narrow, transverse strip adjoining the prothorax and covered at 

 each side by the overhanging gena?. This strip, it seems to the writer, 

 must be all that remains of the base of the occiput, gula, and labium. 



