54 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IO3 



of the posterior plate of the mesonotum (F, J, r), above the posterior 

 end of the first axillary (D). Its narrow outer end is closely pressed 

 against the axillary cord not far behind the proximal end of the third 

 axillary. The fourth axillary is the posterior hinge plate of the wing, 

 but in the bees it has an accessory function in connection with the 

 flexion of the wing by reason of its close association with a leverlike 

 sclerite (D, a.vlv) lying immediately behind it and connected with 

 the arm of the postphragma. 



The axillary lever is an elongate triangular sclerite, as seen from 

 the side (fig. 20 B), lying against the mesal surface of the stalk of 

 the mesothoracic postphragma (A, axlv). By a small process (o) on 

 the ventral angle of its base, directed anteriorly, the lever articulates 

 with the narrowed extremity of the phragma stalk, and on its decurved 

 distal end is attached a relatively large, fan-shaped muscle {y8) aris- 

 ing ventrally on the mesothoracic arm of the pterothoracic endo- 

 sternum. The sclerite is an internal structure except for the end of 

 the thick dorsal angle of its base {s), which comes to the surface in 

 the membrane of the posterior angle of the wing base, and appears 

 here externally as a small sclerite lying immediately behind the 

 fourth axillary (fig. 19 D, axlv). The function of the lever will be 

 described in connection with the mechanism of the wing. 



So far as yet observed the axillary lever is a free sclerite only in 

 the Apoidea. In other Hymenoptera it is represented by an im- 

 movable arm or lobe of the postphragma, on which, however, there 

 is always attached a muscle from the mesothoracic endosternum. 

 In Sphecius (fig. 20 C) and Vespula (D), for example, the lever 

 of Apis is so closely imitated by a long decurved arm {t) arising by 

 its widened anterior end from the mesal surface of the stalk of the 

 phragma that there can be no question of the homology of the two 

 structures. In the ichneumonid Megarhyssa lunator (E) the cor- 

 responding arm {t) is simply a backwardly directed process of the 

 phragma, and in the chalastogastrous genera Sirex (F), Pteronidea 

 (G), and Ciiiibex (H) it is a mere process or lobe (/) of the lateral 

 part of the phragma giving attachment to an endosternal muscle 

 (yS). In these last forms the muscles evidently pull downward on 

 the phragma and probably depress the phragma-bearing postnotal 

 plate {PN2). According to Duncan (1939) the muscle of the phragma 

 arm in Vespula "resists the tendency of the indirect depressor muscle 

 of the forewing (attached on the phragma) to lift the mesopost- 

 phragma, and helps to retract the phragma to the position of rest 

 following a contraction of the indirect depressor muscle." The de- 

 velopment of this structure into an accessory of the wing-flexing ap- 



