BY li. J. TILLYAKD. 



171 



tcance, turning slantingly downwards to run alongside it, and 

 finally leaves Cua by bending downwards and backwards, so as to 

 complete the formation of the distal side of the "anal loop." In 

 the genus J^schna, it bifurcates, near its end, into two small 

 branches. One of these turns basad to join A., thus completing 

 the closure of the anal loop; while the other turns away distad 

 in a curve, and helps to form the well-known secondary loop of 

 this genus (AT, Fig.3). The fate of Aj, in the Libellulince, is 

 also of great interest, and will be dealt with later on. 



Having traced the course and fate of the various branches of 

 A in ^iJschna, we may now exhibit, in a table, the complete com- 

 parison between the two pairs of combined tracheae, viz., R + M 

 and Cu + A. 



In the above table, corresponding parts in the two main 

 columns are true analogues of one another, except in one case, 

 viz., the upper side of the triangle. This is, of course, not 

 analogous to A', but it is included in tlie table in brackets in 

 order to call attention to the manner in which it continues M^ 

 backwards, just as A' continues A4 backwards. 



We turn now to the study of the Anal Loop. The structure 

 known by this name is found throughout the .Eschnin(v. and the 

 whole series of the Libelhdidce. But whereas it varies very 

 little in shape throughout the jEschnincH, it exhibits very great 

 diversity of form in the Libellulidce. On the one hand, we have 



