48 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I03 



with a long slender dilator muscle (z-'j) arising ventrally from a small 

 knob on the pleural margin of the metacoxal cavity just before the 

 pleural coxal process (CxP^). 



The pterothoracic sternal areas. — The ventral surface of the ptero- 

 thorax is entirely sclerotized except in the propodeal region (fig. 

 16 F), and the sternal areas of the mesothorax and the metathorax 

 (^2, S2), though separated from each other by a groove between the 

 bases of the middle legs, are in each segment continuous with the 

 pleural areas (fig. 15). The lateral extent of the mesosternum can 

 be judged only by the areas occupied by the ventral ends of the great 

 pillars of dorsoventral muscles (fig. 16 C, /2), which muscles in 

 generalized insects have their ventral attachments on the sternum; 

 in the bee the bases of these muscles extend well up on the sides of 

 the pleurosternal areas. In the metathorax there are no corresponding 

 tergosternal muscles, but the sternal area of this segment may be 

 regarded as that part of the venter lying between and before the 

 hind legs (F, S^). Both the mesosternal area and the metasternal 

 area are marked each by a median groove, the two grooves being the 

 lines of inflection of a large endosternal structure (H, Endst). The 

 sternum of the propodeum is a narrow, weakly sclerotized transverse 

 band (F, IS) lying in the ventral membrane between the hind legs 

 and the abdominal petiole. Laterally it connects with the lower 

 margins of the propodeal tergum behind the legs. 



According to a recent interpretation of the ventral sclerotization of 

 the insect thorax by Ferris (1940), the apparent sternal areas of the 

 mesothorax and the metathorax of the bee (fig. 16 F, 6*2, S3) would 

 probably be regarded as ventral continuations of the pleura, the true 

 sternal parts being supposed to be entirely folded in to make the 

 base of the endosternum. The evidence presented in favor of this 

 interpretation, however, is argumentative, and does not include a con- 

 sideration of the dorsoventral muscles, which appear to be tergo- 

 sternal in their attachments among generalized insects. It also seems 

 to assume that morphological regions of a segment necessarily follow 

 shifts in the sclerotization, rather than the converse, that sclerotized 

 areas may expand or contract in a continuous and otherwise stationary 

 integument. In the chalastogastrous family Xyelidae, among the 

 Hymenoptera, the episternal sclerotizations of the mesothorax are 

 extended downward onto the ventral surface, where they are con- 

 tinuous across the median line in a narrow or wide bridge. Behind 

 the bridge, however, is a large, triangular basisternal plate with a 

 strong median groove, and on this sternal plate are attached the lower 

 ends of the dorsoventral muscles. In most other Hymenoptera the 



