PHILIP A. MUXZ 21 



50) have few sectors and no branches to Cu-.. Mecistogaster is the 

 most speciaHzed of all, with a single row of cells behind Cu2 and 

 between Mu and M2. It is interesting to note that in this group 

 the tendency is for the nodus to approach a more normal position 

 in the wing, so that in Mecistogaster (fig. 50) it is at one-fourth 

 the length of the wing, while in Megaloprepus (fig. 46) it is at 

 one-seventh. 



The Coenagrioninae 



Baron de Selys created two legions, Platycnemis and Agrion, for 

 the reception of genera which fall, it is true, into two groups, but 

 these have no definite nor sharply defined difl^erences. In general 

 it can be said that the quadrangle of the first of these groups has a 

 relatively long upper side, hence that the outer angle is not very 

 acute, but there are exceptions to this, as for example in Leptocne- 

 mis (fig. 80); in the second legion there areexceptions to the obliq- 

 uity of the quadrangle, since some of the more specialized forms, 

 as Hemiphlebia (fig. 106) and Agriocne?ms (fig. 104), have a fairly 

 regular one. For that reason a recent key to subfamilies of the 

 Zygoptera does not hold as applied in the following quotation: 



"Quadrilateral regular Subfamily Platycneminae. 



Quadrilateral irregular, usually with distal angle sharply acute. 



Subfamily A grioninae.''^ 

 Now these two groups have many points in common; they have no 

 supplementary sectors other than Mia, they have the stigma rather 

 short as compared with that of the IMegapodagrioninae and they 

 have Cu2 well developed. Since they do grade so from one into 

 the other, and are so fundamentally related, I cannot follow the 

 above example and place them in separate subfamilies; in fact, I 

 am not absolutely sure that as many such divisions ought to be 

 recognized as I ha^•e already done in this paper, that is, whether 

 these "subfamily" groups are really equivalent to subfamilies in 

 other orders. 



This placing both in one group does not indicate, however, that 

 my feeling is that the legions Platycnemis and Coenagrion are not 

 rather distinct. Although the quadrangle alone does not suffice in 



^ Agrioninae is here used for the legion Agrion of dc Selys, hence it is equivalent to 

 the Coenagrioninae of this paper minus the Platycnemis group. 



MEM. AM. EXT. SOC, 3. 



