( 347 ) 



II. R. Sphingidae semauophorae. 



(S 'i . The not-scaled area of the iiiiicr surfiice of (lie first segment of the jialpus 

 covered with short sensory liaiis, or tliese liairs, whieli are seldom vestigal,* 

 restricted to a patch. 



The number of 8j)ecies belonging to this division of the Sphiinjidae is about 

 twii'e as large as that of the Sphingidae a^emunophorac. Si)ecialisation by 

 ridnction or loss, though not quite absent, is far less frequent in the present 

 division than among the Acherontiiiiae and Avihulicinae, while the modification of 

 organs in functionally higher structures is very often met with, the ])revailing 

 tendency in the Sphiiigidae semcniop/iorae being progressive development in contra- 

 distinction to the Sphincjidae asemanophorae, where retrogression is the prevalent 

 feature in the development. 



The tongue is never excessively long as it often is in Aclierordiinae, nor is it 

 ever so much reduced as in some Arheroidiiiiae and many Ainbulicinae, reaching 

 always at least to the middle of the abdomen and remaining functional in all 

 species. The jnlifer is always pi'omineut ; it is peculiarly mo<Iified in the 

 Choerocampinae ; the bristles never become weak and flat or change into scales. The 

 palpus varies very much in size, as does also the proportion in length between the 

 first and second segments. In some genera the palpus is somewhat reduced (for 

 instance in ]>eidumia and iJarapsa), in others much enlarged {(Jiganteopolpiis, 

 Elihia, etc.). The longest Sphingid palpus is found in tiiis division, namely in 

 Tinostoma. Remarkable modifications of the palpus occur in the subfamily 

 ('lioerocampinae. The antennae are no less variable. They are filiform or setiform 

 in many species, without the indication of a club ; in others they are moderately 

 clubbed; in others again excessively so. The distal part may be abruptly recurved 

 or scarcely curved at all. The end-segment is more often long than it is short, but 

 the short end-segment is not at all of rare occurrence. The eye is lashed or not ; 

 its size is very diflerent, Onjha having the largest eye of Sphingiilae, while 

 Deidumid, J>arapsa, Gurelca, etc., have a comparatively small one. The head is 

 sometimes crested, the crest being here, as in the Sphingidae asemanophorae, a 

 sign of reduction of the head. The scales of the mesonotum form occasionally also 

 a high, well-defined crest, which may be sim])le (Epistor), or double (Dilophonota). 

 This thoracic crest is not a feature accessory to reduction. The abdomen is of 

 interest in several respects. The first tergite and the sternite of the second segment 

 become often closely appressed to the metathorax (Macroglossutn, Sesia, etc.). 

 The spines are never lost ; there are two kinds of spination not found in the 

 previous subfamilies — the uniserial conical spines of JH/ophonata, Pholus, etc. 

 (PI. LXII. f. 9), and the flattened, very strongly chitinised sjjines of Maeroglossum, 

 Sesia, Haemorrhagia, etc. (PI. LXII. f. II). The spines of the basal sternite, 

 which are at the highest vestigial in Sphingidae asemanophorae, are very distinct 

 in ijuite a number of genera of the present division, and like those of the other 

 sternites barely less strong than the dorsal ones in those forms which have the 

 spination as represented by PI. LXII. f. 11. The broad expansible fan-tail and, in 

 the ?, the broadly trapeziibrm seventh sternite which is spinose at the edge in a 

 number of genera (PI. LXIII. f, .5. G), are characters confined f to members of this 



* See OthmUixida and Giirchci. 



t Excepting the males of the Ambiilicine genera Cypa, SincrittOtidns and DrguKipl, ra, which 

 have also a fan-tail. 



