ATTACIDES. 281 



For the structure of a fairly easily-obtained Attacid pupa reference 

 should be made to our description of that of Saturnia pavonia (posted). 

 Chapman writes : " The Attacid pupa is one that in its more obvious 

 characters is extremely varied, owing to the variety of its environment, 

 some pupating underground, others forming cocoons. In its more 

 essential characters, it is an obtect pupa, very difficult to differentiate 

 from Smerinfhus or Dimorpha, and, at times, from certain Lachneids. 

 It is a thick robust pupa, essentially very round (circular in section) 

 and very straight — Saturnia pavonia and others are obvious ex- 

 ceptions to this, being both flattened from back to front, and with 

 the hinder segments bent well forward. The wings meet in the middle 

 line. In those with well-developed antennae, the greater part of the 

 2nd legs and a portion of the first are covered by them. This 

 occurs also in the Cochlidids and elsewhere. There is no femoral 

 piece between the maxilla and first leg. This is wanting also in 

 Smerinthids, but appears in nearly all other Sphingids. Callosities on 

 the metathoracic dorsum are marked in some Citheroniids, as in 

 Acherontias and other Sphinges, but not so markedly in Smerinthids, 

 they exist in Saturnians, but usually very obscurely. The surface 

 is free from any hairs, such as form a pile on some Lachneids, and 

 is generally smooth and polished; the dorsum, however, in some, e.g., 

 Ant/teraea, etc., possesses patches of brown hairs or rather bristles. 

 In Ant/teraea the pupa is very short and thick, almost spherical, this 

 modification is due to the form of cocoon, and similar modifications 

 go with similar cocoons, in Lachneids, Cochlidids, etc., but it appears 

 to reach its extreme in Antheraea. The anal armature in Cithero- 

 niids generally ends in a double point, and is without bristles. 

 In Saturnians, it is a transverse ridge with, on either side, a row of 

 bristles along its anterior and posterior edge, and somewhat curved 

 over to face each other, the extent to which these are developed varies 

 much, the anterior or posterior row is generally nearly wanting, the 

 bristles are curved, but apparently never actually hooked. In species 

 with cocoons, the segmental armature is rarely more than a little 

 wrinkling, in subterranean species, there is usually a raised row of 

 points along the posterior margin of segments, and a row of pits, 

 or even a raised ridge along the anterior border ; there is no lateral 

 flange as in Enmorpha and higher Sphingids. In Cricula trifenestrata 

 there is a strong set of cremastral hooks clothing a conical projection 

 and not easily interpreted in the terms given above as usually 

 covering the anal armature of Saturnians. In a large African form, 

 very like Ceratocampa, there are callosities or great warts on the pro- 

 and mesothorax as well as on the metathorax, and similar ear-like 

 projections protect the ist spiracles, whilst on the dorsum of the last 

 abdominal segment are two holes leading into large spherical cavities, 

 that look like two peas when seen from within the pupa. The hind- 

 wing is always (?) visible as far as the middle of the hind-margin 

 of forewing. The texture is very hard and thick in subterranean forms 

 and the colour is usually black, whilst it may be delicate and the colour 

 pale brown in forms with solid cocoons, e.g., Cricula, which has a fairly 

 solid texture. It may be a question whether any absolute characters can 

 be given for the pupae of any of the families of larger moths with obtect 

 pupae, many of which are closely related, and others closely approximated 

 by convergence, whilst in each family there is much variation of an 



