NO. 1389. DRAGONFLIES OF BURMA AND SIAM— WILLIAMSON. 167 



De Sel3^s^s statement, since it is not shown that he, Laidlaw himself, 

 did not overlook the females. His failure to discover them may have 

 been because of one or both of two possible reasons: They may con- 

 ceal themselves in the veg-etation near the haunts of the males, appear- 

 ing- on the wing rarel}" and then at hours of the day not suspicioned 

 by the dragonfl}^ collector; or they may habitually frequent retreats 

 in the jungles far from the localities where the males are conspicuous 

 with l)rilliant wings to attract the collector to the most favorable 

 hunting g-rounds. There may be other reasons, too, wh}" the females 

 are less known in collections, but the a])ove remarks will indicate that 

 orig'inal observations are possible and desirable. In the present paper 

 only imagoes are discussed. Nothing could be more desirable in this 

 connection than a collection of nymphs sufficient to at least allow of 

 the delinition of generic characters in immature stages. 



KEY TO THE ORIENTAL GENERA OP THE SUBFAMILY CALOPTERYGIN.E (lMAGOEs).« 



ORDER ODONATA (Neuroptera Odonata, Paraneuroptera). 



a. Fore ami hind wings similar in veiiational structure; quadrangle i)resent; wing 



meniliranule wanting; supplements (veins across wing mendjrane opposed to 



principal veins) w'anting. Head wide, eyes separated. Males with 2 inferior 



abdominal appendages; females with genital valves. . .Suborder ZYCjOPTERA 



h. Mj separating from M]+., nearer the nodus than the arculus. 



Family Agrionid.e 

 hh. M;, separating from M1+2 nearer the arculus than the nodus, or at not more 



than half the distance from arculus to nodus Family Calopterygid.e 



c. Antenodals 2 subfamily lestin.ic 



cc. Antenodals 4 or more subfamily calopterycjix.e 



d. M1-3 and M^ rising from the middle or near the middle of the arculus, not 

 together from its extreme anterior end. 

 e. Antenodals of first series not coinciding with those of second series beyond 

 the level of the arculus; M1+2 and M3 forming a symmetrical fork, or 

 Ml +2 continuing the direction of Mj-.j; wings petioled at least one- 

 half way to the arculus; sul)quadrangle bent at the arculus. 

 /. Antenodals of second series wanting beyond the level of the arculus; 

 normally 4 antenodals continuous in the first and second series; 

 wings petioled and narrowed nearly to the level of the arculus and 

 half way to the level of the nodus; distance from base of front wing 

 to nodus less than J wing length'; M2 arising beyond the nodus at 

 about the fourth postnodal of the second series; quadrangle about ^ 

 as long as basal space and with 2 cross veins; M,.^ ]xirallel to ^l^; 



stigma ovate Devadatla 



ff. Antenodals of second series present beyond the arculus; wings less 

 petioled and the nodus not so retracted. 

 g. Second antenodal over arculus; Mj-jand M^ arising from a single 

 point; quadrangle wdder proximally MieromerKs 



«Two genera, Calipliixa and Philoganga, known to me only from descriptions, are 

 omitted. Neither genus is known from Burma or Siam. 



