GENUS CASTNIA AND SOME ALLIED GROUPS. 161 



arrangement as compared with the simple system seen in so many other groups of 

 Lepidoptera. The numher of branches is normally present ; but the 2nd, 3rd, 4th, 

 and 5th branches of the postcostal (b 2, b 3, b 4, and b 5) are generally arranged so that 

 they form a small oblong premedian or postcostal cell. Looking at PI. XXVIII. fig. 6 

 (O. cacica), we see that this is effected by the branch b 3 throwing off a very short 

 transverse branchlet connecting it with b 4 ; or, rather, we may say that b 4 commences 

 transversely, and then suddenly bends at a right angle, and runs longitudinally, so that 

 the little transverse branchlet must be regarded as its basal portion. In b 5 we have 

 the anomaly that the 5th branch arises at a short distance beyond the 1st branch, and 

 forms the hinder margin of the postcostal cell, throwing off a little oblique branchlet, 

 by which it is united to the 4th branch, and which in fact forms part of the extremity 

 of the postcostal cell. In like manner the median vein (c) has its third branch (c 3) 

 strongly angulated at a right angle at its base, whilst the branch c 3* (lower discoidal 

 nervule of Doubleday) is carried backwards by a fold (z), the discoidal nervure of 

 Doubleday, and almost to the base of the wing f . Moreover the veinlet connecting b 5* 

 and c 3* is angulated below its middle, throwing off backwards a fold which unites 

 with 3, and which indicates the normal position of the discoidal nervure of Doubleday. 

 The object of this complicated arrangement, which has not hitherto been described, 

 appears to be to form a strong network of corneous veins for strengthening the wing, 

 so as to enable it to perform the strong flight which these muscles doubtless enable 

 the insect to perform. 



Modifications of the arrangement described above exist in many of the species of 

 Castnia. 



In the splendid new species C. veragiiana (fore wing, PI. XXVIII. fig. 9) the distribution 

 is nearly similar to that of C. cacica ; but there is no little transverse branchlet connecting 

 b 3 and b 4, and consequently there is no closed postcostal cell, and the branch c 3? arises 

 from c 3 in the ordinary manner at a little distance in front of the transverse base of 

 c 3, making an acute angle instead of being pushed backwards a little distance, and 

 making an obtuse angle as in O. cacica. The curious species C. cochrus (PL XXVIII. 

 fig. 11) agrees with C. veragiiana in this respect. 



In C. Boisduvallii and other closely allied species (PI. XXVIII. fig. 8) the small narrow 

 postcostal cell exists ; but it only emits one branch at its extremity, which branches off 

 into 6 3, 6 4, and b 5, and the angulated base of b 5* arises from the posterior margin 

 of the postcostal cell near its extremity. 



In C. icarus (PI. XXVIII. fig. 13) the small postcostal cell exists, emitting b 2 from its 

 front margin a little beyond the middle, whilst from its extremity one branch is emitted 

 which forks into b 3 and b 4, whilst b 5 arises from the extremity of the postcostal cell, 

 close to the base of preceding branch, and b 5* from b 5 at a little distance from its base. 



In C. Subneri (PL XXVIII. fig. 14) there is no closed small postcostal cell, the little 

 transverse veinlet between b 3 and b 4 being wanting. In Gray's figure, pi. 13. f. 2, 

 the vein is misrepresented as closing the cell. 



In C. Latreillii{J?\. XXVIII. fig. 15) the small postcostal cell exists, terminating in an 



t G. linus scarcely differs from C. cacica in the arrangement of the veins of its fore wings (PI. 5XIX. fig. 9). 



