THE PHYLOGENY OF THE LEPIDOPTEROUS PUPA. 99 



The Tortricids may be taken as presenting very typical pupas- 

 incompletse. Throughout the whole superfamily there is very great 

 uniformity of character. The first three segments of the abdomen are 

 fixed and attached to the wings, but not so firmly but that, on 

 dehiscence, the incision anterior to the third abdominal opens dorsally 

 more or less. Two rows of hooks extend across the dorsum, an 

 anterior and posterior on the 2nd and following abdominal seg- 

 ments, but so close together that, in a dehisced pupa, the space 

 between them may not exceed one-fourth the width of the segment. 

 On the 8th and 9th abdominal segments they are reduced to one row 

 (the anterior), and on the 10th there is usually a cremaster with 

 hooked hairs or bristles. The spines, even when nearly obsolete, are 

 always represented by a transverse ridge. The maxillae are separated 

 widely enough basally to show nearly all the labial palpi, and a 

 maxillary palpus is sometimes actually or very nearly wanting. A 

 dehisced Tortricid pupa is a most beautiful and instructive object, but 

 difficult to handle by way of description. The whole pupa is arched 

 by the extension of all the abdominal incisions, even that between the 

 1st and 2nd abdominals being often opened a little. The dorsal slit 

 extends through the whole thorax. The head and antenna? are lifted 

 forwards, the antennae quitting their groove between the wings and 

 legs. The first two legs, including the 1st femur, form the two pieces 

 between this and the wings. At the front of the prothorax, which 

 opens a little from the mesothorax, is a small plate, the dorsal head- 

 piece. Forwards from this, and connected with it by a film that 

 passes under the antenna, is the eye-cover. These parts are all sus- 

 tained in their places by the pupal skin of the covered parts. From 

 under the face-piece a film extends backwards, formed of the front 

 covering of the thoracic, and first three abdominal, segments, and, from 

 each side of this, the coverings of the trochanters, femora, &c, form 

 supports for, and connections with, the leg-pieces. 



The distinction between the Sesiid and Tortricid pupae is very narrow. 

 The Sesiid is very close to, if really distinguishable from, that of a 

 true Tinea. The Sesiid pupa has always anal spines arranged more or 

 less as a circle or coronet, never a true cremaster ; the Tortricid pupa, 

 when without a cremaster, is here indistinguishable. In the Sesiid pupa, 

 the wing-cases project beyond the 3rd abdominal, which is never the 

 case in the Tortricid pupa, though it is very probable that there will be 

 found to be exceptions here, as possibly in another of these characters, 

 which all mark Sesia as a less generalised form than Tortrix. This is 

 in the dehiscence, the Tortricid pupa retaining the three leg-covers of 

 each side (1st femur, 1st leg, 2nd leg) in one piece, whilst in the 

 Tineid and Sesiid pupae they distinctly separate. The maxillary 

 palpus is well developed in Sesia. 



The pupa of Zeuzera has a somewhat intermediate position, the 

 femoral piece separates, but not the two legs ; there is no cremaster, but 

 a provision of ventral spines ; the wings separate from the abdominal 

 segments on dehiscence, but are previously loosely attached to the 1st 

 and 2nd abdominals, and extend to the end of the 3rd. It agrees 

 with the Sesiid rather than the Tortricid pupa in the full development 

 of the maxillary palpus. This is not, however, distinctive, since many 

 Tortricid pupae have well-developed maxillary palpi, such, for instance, 

 as Semasia woeberana, which has, however, the femur attached to the legs 



