364 BRITISH LEPIDOPTERA. 



Larva.-— (i) Colour of ist stage. (2) Tubercles bearing more than one 

 hair (1st stage). (3^ Habits (as a whole). (4) Colour of shagreen tubercles (2nd 

 stage). (5) Shape of head (2nd stage). (6) Additional stripe on horn or hump of 

 the 8th abdominal (best seen in last stage). (7) Double row of black spots (dorsal) 

 2nd stage. 



Pupa. — (8) Bristles on anal armature : as in Satumia pavonia. 



Dyar concludes {Ann. New York Acad. Set., viii., p. 232) that the 

 affinities of the larvae suggest a relationship , between Sphingids, 

 Notodontids and Lachneids, an impossible combination, but his 

 description of the larva here is so faulty that further criticism is need- 

 less. Later he gives (Trans. New York Acad. Set., xiv., p. 59) the 

 following accurate diagnosis of the Sphingid larva : 



The tubercles all remote, v moved up * in front of the spiracle, all the seta? 

 disappearing or becoming obscured at the first moult ; tubercles i on the 8th 

 abdominal segment borne on the apex of a long process (caudal horn), but they are 

 entirely unconsolidated. 



He then says : " The consideration of the first larval stage shows 

 plainly that the Sphingids are not related to the Attacids, but 

 rank as a separate division." He further notes that Poulton's 

 supposition that the thoracic horns of Ceratomia amyntor might be 

 homologous with those of the Cither oniidae, can be shown to be 

 unfounded, for, in the first larval stage of this Sphingid, these pro- 

 cesses arise anteriorly to the setae and entirely independently of them, 

 whereas in the Citheroniidae the horns are developed out of the 

 corresponding tubercles. Dyar's suggestion of a relationship with 

 the Notodontids is contradicted on all other characters than those 

 he notes — egg, pupa, imago, and even, in their broader aspects, by 

 those of the larvae. 



Meyrick's note on the phylogeny of the group (Handbook, etc., 

 p. 293) is characteristically airy. He writes : " The phylogeny is 

 sufficiently simple ; the group of Smerinthus and Dilina is of primitive 

 character, and some exotic members of it closely approach the 

 Notodontidae ; the other genera constitute a more largely developed 

 line of descent originating in the group." We wish we could accept 

 this simple statement. The Sphingids (sens, lat.) are proved by the 

 egg, larval, pupal and imaginal stages to have no structural relation- 

 ship whatever with the Notodontids, and, certainly in the larval 

 and some imaginal structures, the Amorphids (Smerinthus and 

 Dilina of Meyrick) are probably more specialised than any other 

 Sphingids. The evolution within the group is not simple. It is quite 

 clear, as we have already pointed out, that the primitive Sphingid 

 had, among others, the following characters : 



Ovum : Oval, transparent shell. Larva : Primitive tubercles from meso- 

 thorax to 7th abdominal ; hump bearing i on 8th. PUPA,: Rough skin, with well- 

 developed dorsal points or spiues. Imago : Well-developed (not necessarily long, 

 but functional) tongue, frenulum, two anal nervures to hindwings. 



As a matter of fact one finds within the group: (1) 

 A very general and characteristic form of egg with scarcely 

 any modification. (2) Larvae represented by the Eumorphids 



* We doubt this ; we believe v to be atrophied and the prespiracular here 

 called v to be quite distinct from v. 



