34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8o 



ment flattened out in the lateral wall of the body segment. In some 

 holometabolous larvse, as in the larva of a carabid beetle (figs. l6 A, 

 25), or in a sawfly larva (fig. 16 B), the thoracic legs are carried on 

 subcoxal mounds (Sex) of the body wall, which have the appearance 

 of being the true bases of the legs, and in which are situated laterally 

 the pleural sclerites. 



To trace the evidence of subcoxal elements in the abdominal seg- 

 ments of adult and larval insects (fig. 16 C) would take us beyond 

 the limits of a paper on the thorax, but the contention by Borner 

 (1909) that the gills of mayfly nymphs are homologues of the tho- 

 racic legs, and that their basal supports are subcoxae is too interesting 

 to pass over. Diirken (1909), it is true, has shown that the muscu- 

 lature of the gill does not fit with Borner's interpretation ; but if 

 Borner had compared the gill muscles, not with the tergal promotors 

 and remotors of the coxae, but with the coxal abductors and ad- 

 ductors, which should arise within the subcoxae, there might be less 

 objection to his homology of the skeletal parts. The general subject 

 of the musculature of the leg base, however, will be given special 

 attention in a following section (page 83), wherein it will be shown 

 that the coxa appears to possess muscles that should belong to two 

 leg segments. 



Borner rightly says that an arthropod limb, in order to be an effec- 

 tive instrument of locomotion, must be able to turn forward and 

 backward on its base. If a subcoxal segment was the primitive base 

 of the limb, it, therefore, moved upon a vertical axis in the pleural 

 membrane between tergum and sternum. Borner assumes also that 

 the coxa moved in the same manner as the subcoxa; but this would 

 give a double- jointed movement of the limb base in one plane. More 

 reasonable does it seem that the primitive coxa moved in a vertical 

 plane on a horizontal axis between anterior and posterior articular 

 points on the subcoxa, thus duplicating the movement of the follow- 

 ing coxo-trochantinal joint, rather than that of the limb with the body. 

 When the subcoxa then became a fixed part of the body wall, its 

 function as the leg base necessarily devolved upon the coxa, and the 

 latter, taking over the promotor and remotor muscles of the subcoxa, 

 shifted its axis from the horizontal to an oblique position and finally 

 to a vertical one. 



The subcoxal theory, then, as proposed in this paper, assumes that 

 the coxa originally articulated with the hypothetical basal segment of 

 the limb (fig. 18 A, Sex) by an anterior articulation (a) and a pos- 

 terior articulation (b), and that, in insects, the lateral walls of the 

 subcoxa have furnished the pleural sclerites of the thorax, while the 



