MANDUCA ATROPOS. 399 



amount of regular alignment affects the other scales, but even then 

 it is not easy to say that they fall into two other rows. The male 

 antenna is of about equal thickness, romm., from the ioth to the 

 35th joint, the female is only about o'jmm. in thickness and does 

 not reach this till about the 13th or 15th joint, the terminal bend 

 and tassel consist of about 15 joints. The fore tibia and its spur 

 are approximately identical in the two sexes; the chitinous surfaces 

 of these first tibial spurs are covered with fine spicules, varying 

 from the spines of the comb to hair-like and scale-like structures, 

 which are very beautiful. The male scent fan* is very well developed, 

 it arises from the margin of the ventral plate of the 1st (2nd)f 

 abdominal segment over a length of rather over a millimetre, it is 

 very dense, composed of reddish-orange hairs, 5mm.-6 # 5mm. in length. 

 These are concealed in an unusually large pocket, in its own 

 and the following segment, consisting of a deep groove, formed by 

 a depression of the margin of the ventral plate, which is covered by 

 a reflexion inwards of the margin of the dorsal plate. The palpi of 

 M. atrofios have a structure which is almost exactly the same 

 as is found in A. convolvuli, and which I find also in A. cingulata 

 and another closely related Agriid. It appears to be present in 

 the species of Manduca, but I cannot find it in any other Sphingid 

 species that I have been able to examine. The inner face of 

 the 2nd joint in M. atropos has a central basal portion, convex 

 and unclothed, this occupies about f of the length ; beyond this is a 

 rather deep circular hollow, from the basal margin of which arises 

 a row of rather flat scales which arch over the hollow, forming a 

 dome-like roof over the hollow, and which meet together at its further 

 margin. It is this fan-like dome that is seen on looking at the inner 

 surface of the palpus. In A. convolvuli this dome of scales is even 

 more beautifully elaborate. It arises nearer the base of the joint, 

 the portion basal to it is scaled and slopes down to it. The origin 

 of the fan-like dome of scales is of very regular circular outline, as 

 is the whole dome, although the scales converge to a point at the 

 concave (upper) margin of this joint, the hollow beneath the dome 

 extends beyond it to the end of the joint and is thus not circular 

 (though the dome is). The dome of scales thus occupies about the 

 centre of the inner face of the joint, but rather nearer base. The 

 hollow is unclothed (Chapman). 



Gynandromorphism. — Only one gynandromorph of this species 

 is known to us. This is described as follows : 



a. The left side in comparison with the right betokens the ? sex in that it 

 has a somewhat smaller antenna, and a shortened abdominal point ; otherwise 

 the specimen is <? (Taschenberg, Zeits. f. d. ges. Naturwissenschaften, xxii., 

 pp. 520-521). 



Variation. — The variation of this species has sometimes quite 

 a teratological basis. Such variation as that of the shape of the 



* Bartel notes (Pal. Gross- Schmett., ii., p. 22) : "In the <? , on each side of 

 the 1st abdominal segment, is seen a deep pocket, from which, in flight, a wdiole 

 fan of long-, strongly-scented, yellow hairs can be thrown. When the S s have 

 been killed, this scent-apparatus is only very rarely visible. One can, however, 

 see the verticillate tuft of hairs if one so holds a living s by the abdomen, that the 

 underside of the abdomen is turned upwards and the wings can be moved." 



t We have previously noted that the 1st abdominal segment is nearly obsolete 

 and that the apparent 1st is really the 2nd segment. 



