NO. 1687. THE THORAX OF INSECTS— f^NODGRASS. 519 



Now, it is only in the mesothorax and metathorax of Mecoptera 

 Neuroptera, Trichoptera, and Lepidoptera that the coxa is actually 

 a double structure. In these orders the coxa genuina of Walton 

 carries the trochanter, while the coxa meron is attached to the coxa 

 genuina only. In other orders in which the coxa shows a more or 

 less evident division this division is in the coxa genuina itself, the 

 coxa nieron being absent, and is of the nature of a strengthening of 

 the coxa by opposite ridges on its inner walls. In the Neuroptera 

 and Trichoptera at least it can easily be demonstrated, by a study of 

 larval and pupal forms, that the " coxa meron " is simply a detached 

 extension of the epimerum, which fuses upon the posterior side of the 

 true coxa. It is, hence, not a part of the primitive coxa at all, and 

 the apparent double coxa in these orders is a purely secondary con- 

 dition. (See special descriptions under Neuroptera, p. 5G4, and Tri- 

 choptera, p. 565, also p. 542 and figs. 144-148.) 



Comstock and Kochi (1902) show that the plates of each segment 

 may be arranged into two subsegments, but defer any opinion as to 

 whether they represent two primitive segments or not. 



It Avill be found that all these theories are purely imaginative, 

 Embryologists have not shown that the plates of any thoracic seg- 

 ment are derived from more than one metamere. However, it may 

 be true that two, three, or four segments primarily existed where 

 there is but one in insects as we now know them. Verhoeff (1902, 

 1903, 1903a, 1903c, 1904, 1904a) is the principal elaborator of this 

 theory. 



Verhoeff bases his ideas on a study of the Aptera, the Embiidse, 

 and the Dermaptera, and especially on a comparison of Javyx with 

 the Chilopoda. He first points out the tendency in the Chilopoda 

 toward the suppression of every alternate segment by a fusion with 

 the preceding larger spiracle-bearing segment. In Japyx there are 

 remnants of extra segments between the pro- and mesothorax, and 

 between the meso- and metathorax, represented principally by well- 

 developed tergal and sternal plates. Thus the thorax would consist 

 of six segments in three pairs, namely, the microthorax and prothorax, 

 the stenothorax and mesothorax, and the cryptothorax and meta- 

 thorax. Verhoeff observes, however, that this arrangement does not 

 correspond with that of the Chilopoda, because the small segment 

 in Japyx is associated with the large segment following instead of 

 with the one preceding. Then, as if to remedy this discrepancy, he 

 further discovers traces of still other thoracic segments, one between 

 the stenothorax and the mesothorax and another between the crypto- 

 thorax and the metathorax. Finally, by the aid of small presternal 

 plates (" vorplatten ") he is able to construct the following table of 

 complete uniformity in segmentation between Scolopendrida? and 

 Japygidse (Verhoeff, 1904a) : 



