No. 1774. THORAX OF FfYMENOPTERA—SNODGRASl^. 63 



SO as to overlap the segment behind. The second process has taken 

 place m those segments of all species that have a postphragma, 

 including many that have only a rudimentary phragma, and has 

 resulted in the formation of the postnotum wherever this sclerite 

 occurs. The third process may be supposed to have taken place in 

 the mxcsothorax of those orders that have no postphragma and no 

 postnotum in this segment, and in which the notum lies close to that 

 of the succeeding segment, if indeed it is to be assumed that this 

 condition is primitive in such cases and not secondary. The reduc- 

 tion or absence of the postnotum in the metathorax is, of course, a 

 secondary modification consequent upon the reduction of the hind 

 wings.^ The anterior phragma and the posterior phragma are con- 

 stant in their attachment to the front of the mesotergum and the 

 back of the metatergum, respectively, while the middle phragma is 

 assigned to the segment most in need of it. When the front wings 

 are used in flight as much as the others or more the middle phragma 

 is attached to the mesotergum ; when the hind wings are the chief 

 organs of flight it is attached to the metatergum. Thus it results 

 that the principal flight segment is always provided with both a 

 prephragma and a postphragma, while the other is left with only a 

 prephragma or a postphragma. In this way the longitudinal 

 muscles of this favored segment are enabled to act most forcibly on 

 the tergum, though at the expense of some of the power of the 

 muscles of the other segment. In the higher Hymenoptera this 

 specialization has been carried so far that the metathoracic muscles 

 are rudimentary, while the great mesothoracic mass of muscles effects 

 the thorax as a whole, producing the motion of both pairs of wings. 

 Thus it is possible to see a reason for the fundamental structureof 

 the wing-bearing thoracic terga, a structure which follows logically 

 from the assumption that the flight function has been secondarily 

 acquired, and that extra parts had to be added to the primitive notal 

 plates to enable the longitudinal muscles to depress the wings by ele- 

 vating the notal plates, instead of pulling the segments together, 

 which latter was their original function. Furthermore, the strain of 

 these muscles on the notum must be held partly responsible for the mod- 

 ifications of this plate. However, since the function of elevating the 

 wings devolved upon the primitive vertical muscles of the mesothorax 

 and metathorax, it can not be doubted that the primary cause of the 

 modifications of the notal plates is to be traced to this latter source. 

 The foregoing is a brief review of a subject that might be studied 

 and illustrated in much greater detail. The basis of the writer's 

 mformation is contained in his former paper (1909) on the thorax 

 of msects, in which, however, he would now make certain modi- 

 fications mentioned in the present paper. It is hoped that enough 

 new material is given here, first, to substantiate the claim that each 

 wing-bearing tergum of the insect thorax is not composed of four con- 



