March, 1902] 



PSYCHE. 



323 



is a little longer than in Attacus and the 

 other veins of the wings are as in Atta- 

 cus. The venation of the hind wings is 

 nearly the same in both genera. The 

 wonderful similarity of markings, espe- 

 cially the large, clear discal spots in gen- 

 era quite remote is an interesting case of 

 convergence. 



The larva of Rothschildia approaches 

 Samia, rather than Attacus. That of A. 

 atlas has been well described and care- 

 fully figured in all stages by M. Poujade, 

 (Annales Sec. Ent. France, X, 1880; p. 

 183, PI. S.) 



The larva of Attacus atlas in its final 

 stage is provided with long finger-shaped 

 tubercles; those, however, on the ter- 

 gum of the 2d and 3d thoracic segments 

 are very different in shape, being large, 

 short and rounded, those on the abdom- 

 inal segments long and slender. Re- 

 duction occurs on the thoracic segments 

 only ; the two rows of tubercles on the 

 sides of the thoracic segments are of the 

 same shape, but a little longer than those 

 on the abdominal segments. 



In Rothschildia, as shown by blown 

 examples of Rothschildia orizaba received 

 from Mexico, the tubercles are more ru- 

 dimentary ; they are low, short, fleshy, 

 and are crowned with 5-7 small sharp 

 spinules, while those of Attacus atlas are 

 long, finger-shaped and unarmed with 

 any spinules. The median tubercle on 

 the 8th abdominal segment is very small, 

 inconspicuous, and but slightly larger 

 than the other dorsal tubercles of the 

 abdominal segments. The dorsal tu- 

 bercles on the meso- and metathoracic 



segments are scarcely larger, if any, than 

 those in the abdominal segments. 



Burmeister has figured the larvae of 

 Rothschildia hesperus, cthfa, aiirota, betis, 

 and speculifer. In all except R. betis, 

 they agree well with the larva of R. Ori- 

 zaba \ the thoracic dorsal tubercles being 

 no larger than the abdominal ones, this 

 species approaching nearest to R. auro- 

 ta. In R. betis, however, no traces of 

 tubercles are given, and in the text it is 

 stated that the larva has no spines ; the 

 larva is blackish, banded transversely 

 with deep pink-red. The larvae, then, 

 of the American species hitherto re- 

 ferred to Attacus, appear to present ex- 

 cellent distinctive characters. 



Judging by the larvae, whose tubercles 

 are more like those of Samia, Roths- 

 childia is the more primitive type, and 

 Attacus the more specialized. Attacus 

 is in venation and the shape of the wings 

 closely allied to Philosamia {P. cyiithia) 

 its larva is more specialized than that of 

 Philosamia ; but the latter has begun 

 to be specialized in the reduction of the 

 dorsal tubercles of the prothoracic seg- 

 ment, which are short, rounded, and un- 

 armed. 



Attacus is confined to the Oriental 

 region, while the older more primitive 

 genus Philosamia is represented in equa- 

 torial Africa as well as the East Indies ; 

 it is probable that the Ethiopian realm 

 was the original home of these two gen- 

 era, unless Attacus separated after mi- 

 gration into the East Indies, India and 

 the East Indian Archipelago. 



