EUMORPHIN^E. 43 



ments of certain Manducid and Sphingid larvae. The differences 

 between the Eumorphid (sens, lot.) and Sesiid larvae are possibly- 

 best shown in the tabulation given herewith. 



The habits of the Eumorphid larvae are most diverse. The 

 swimming ability of the larva of Eumorpha elpenor will be 

 referred to in our account of this species. Young notes the 

 occurrence of many larvae of a non-British Eumorphid on a 

 species of water-ranunculus at Mahdopoor in the Punjaub, which 

 he came across while up to his knees in water snipe-shooting. 

 Most of the Eumorphid larvae feed on low-growing plants, a few 

 only on shrubs. 



The Eumorphid pupa is usually ot a pale colour, and, 

 in this respect, differs from those of other Sphingids. It is 

 to be observed that the pupal structure appears to be some- 

 what at variance with the larval, for, whilst the larval pecu- 

 liarities give a distinct line of cleavage between the Eumorphid 

 and Sphingid larvae (sens, strict.), the Eumorphid pupal characters are 

 definitely Sphingid (i.e., not Amorphid). Chapman notes : The 

 pupae of the species examined, without being of delicate texture, 

 differ from those of the other Sphingid groups in having a terra- 

 cotta-colour, instead of the brown chitinous colour of Sphinx and 

 Manduca, and that is familiar to us as that of underground pupae, 

 such as those of Noctuids. The terra-cotta may be much overlaid 

 with black markings. It probably has relation to a habit of not 

 going deeply underground for pupation, but of making a cocoon 

 immediately below the surface, amongst rubbish, especially debris 

 of dead leaves, &c, or, if a strict line be drawn, between solid 

 earth and surface debris, not really burying themselves at all. We 

 may recognise several types amongst them, though they are much 

 alike in their general outline, the following being illustrative : 



i. This pupa may be taken as the least specialised, it is smooth, i.e., free 

 rom any spines or flanges, but is very definitely sculptured with a wrinkled surface, 

 the labium is at the anterior extremity of pupa —Hyles euphorbiae, Celerio gallii. 



2. This differs little from the former, except in one important particular, the 

 development on abdominal segments 5, 6 and 7, of a row of small sharp spines, 

 dorsally strong and towards anterior margin of segment, but laterally getting 

 weaker, passing behind the spiracle, and absent ventrally. [The lateral position 

 of these [behind the spiracle) shows they have nothing to do with the characteristic 

 prespiracular flange of Sphinx nor with the anterior rows of spines, common in 

 " pupse-incompletae." I have not myself seen an Eumorphid pupa with the pre- 

 spiracular flange well-developed.] — Eumorpha elpenor, Theretra porcellus. 



3. This pupa is much smoother than that of Hyles [euphorbiae), the wrinkles 

 being very flat, or almost obsolete, the head very large, the carina of proboscis 

 very strong, passing right round the front of the head, and making the labrum 

 absolutely dorsal — Isoples alecto, Hippotion celerio, Daphnis nerii. 



4. This presents another form, differing in the anal armature and in 

 possessing a trace of the prespiracular flange, and in the proboscis having, not 

 merely a strong anterior flange but, a long horn passing directly forwards and 

 returning in a curving sweep, and dipping into and soldered to the ventral portion 

 of the proboscis at the level of the prothorax — Panacra vigil. 



It may be observed that Thaumas vespertilio has a pupa very little 

 different from Hyles euphorbiae, suggesting that, in this case, the colour- 

 ing of the larva indicates its affinities more truly than its want of horn. 

 Considering how close many of the pupae are, T. vespertilio might 

 be pupally characterised as a separate genus, by having the wrinkling 

 so modified that the pupa might almost be described as having 



