542 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxxvi. 



The coxa (Cx) is too familiar to need any special description. As 

 already shown (text fig. 3) it is articulated dorsally to the coxal 

 process of the pleurum {CxP) and ventrally to the coxal process of 

 the trochantin (TnC). The latter may be a sternal element, but it is 

 only in the nymphs of Odonata that the coxa is articulated directl}^ 

 to the sternum (11, 10, d). In the Hemiptera and the Coleoptera the 

 upper ends of the coxae are usually hidden b}^ the overlapping pleu- 

 rites, but even in such cases the coxa will usually be found articulated 

 to a hidden coxal process of the pleural ridge. 



The coxa has often been conceived of as a double structure i-epre- 

 senting elements corresponding with the episternum and the epi- 

 merum. The basis for this idea is furnished chiefly by the Xenroptera, 

 Trichoptera, and Lepidoptera. Here the meso- and motathoracic 

 coxte of adults are composed distinctly of an anterior and a posterior 

 segment (147, U8, 140, 153, Cx and cpm). Banks (1893) regarded 

 the coxa as derived from two segments, tlie leg being the appendage 

 of the first in each pair and the coxal spur of M<i<J(Uis representing 

 the appendage of the second segment. Walton (1!)00) described the 

 two coxal segments as the " coxa genuina " and the " coxa meron." 

 Packard (1808) called the posterior division the 'trochantin." 



A study of larval and pupal forms in the Xeuroptera and Trichop- 

 tera shows that tliis double structure of the coxa is a purelj^ secondary 

 condition. In the larva of Coryddlis vornuta (144) the meso- and meta- 

 coxa^ are simple structures like the prothoracic coxa. The epiiuernm is 

 divided by an oblique groove into an upper })late {Ep/x) and a lower 

 plate [cpin). In the pupa ( 14."») the lower epimeral subdivision (cpm) 

 has extended downward behind the coxa (Cx) and is partially joined 

 to it. In the adult (147) the lower epimeral plate (epm) is entirely 

 fused upon the rear side of the coxa (Cx) and, moreover, is separated 

 by a membraneous area from the upper epimeral plate (Epm). This 

 Avould appear effectively to dispose of the bisegmental notion of the 

 structure of the coxa in this order. The same devel()i)niental process 

 can be shown to take place in the Trichoptera. A })upa of Neuronid 

 ocellifera (140) has a long extension {epm) of the epiuierum {Epm) 

 united to the posterior edge of the coxa {Cx). In the adult (148) this 

 is separated from the upper part of the epimerum and appears as a 

 posterior segment {epm) of the coxa {Cx). Probably the same con- 

 dition could be shown in a freshly pupated Lopidopteran. 



Hence, it may be concluded that the doul)le structure of the coxa 

 in these orders is purely a secondary modification and can in no way 

 be used as evidence of a bisegmental origin of the thoracic segments. 



.'!. TlIK 8TKHNUM. 



The writer has not been able to make an extensive study of the 

 sternum. Comstock and Kochi (1002) have recognized three trans- 



