NO. 1389. DRAGONFLIES OF BURMA AND SIAM— WILLIAMSON. 177 



the first row coii.sisting of "2 spots, the anterior spot one-cell wide, 

 lying- between R^ and M3, the posterior reaching from M^ to the hind 

 margin of the wing. The second row consists of 3 distinct spots, the 

 anterior lies between Mj and M^, the middle spot l)etween R^ and M^, 

 and the posterior between M^ and Cui; an imperfect spot is present 

 between the middle and posterior spots; undoubtedly in some indi- 

 viduals this disappers and in others a well-marked spot is present, 

 variable as in the case of fenestrella. The third and apical row con- 

 sists of a single large spot lying between Mja find the first secondary 

 sector anterior to M.,. Stigma of front wings black, median portion 

 subapicallv yellowish brown, not sharpl}' dehned; of hind wings dor- 

 sally black, with nearly half subapically pale yellow isli brown, ven- 

 trally with the pale area more sharply defined, pale yellow in color. 

 Abdomen black. 



Distinguished from fenestrella ))y having the anterior of the median 

 row of spots decidedly nearer the nodus (beginning before Mja), with 

 the result that the spot is narrower, and its anterior edge is more 

 bounded by M^, while in fenestrella the basal anterior ])order of this 

 spot is more or less formed by Mja- In feriestt'eTla^ moreover, the three 

 spots of the median row are about equal in length, while in qiiad- 

 rimaculata. the middle spot is abbreviated basally a])out half the 

 length of the other two spots. The large apical spot is, in quad- 

 rimacidata, not entirely under the stigma as in fenestrella^ and it is 

 decidedly larger, although this last point will be found undoubtedly 

 subject to great variation since the anterior and postei'ior borders of this 

 spot in the specimen l)efore me are formed h\ secondary sectors in both 

 cases; for this apical spot to increase or decrease one cell's width either 

 anteriorly or posteriorly should not l)e unexpected. Differences in 

 the stigma also exist between /J';;r.sv'/v.//rt and quadrhn.aculata as pointed 

 out in the ke}^ above. In quadrirnacidata the yellow stripe on the 

 metepimeron is longer than in fenestrella and more continuously 

 parallel to the latero-ventral metathoracic carina. QaadrlmacuhUa 

 has hitherto been descri])ed as larger i\\'M\fenestr(dla\ the now known 

 range in size is abdomen l!»-28 nun., hind wing 21-25. I believe 

 qxadriniacnlata is specifically distinct from fenestrella, but indistin- 

 guishable from spuria. 



