MORPHOLOGY OF THE LEPIDOPTEEOUS PUPA. ' 195 



Part II. — The Number of Abdominal Segments and their Relation to 



THOSE of the Larva. 

 It is obviously necessary to determine these points before proceeding to the consideration 

 of the external organs of reproduction, for the segmental relations of the latter are of the 

 highest importance. 



1. The Number of Abdominal Segments in the Larva.— There is no difficulty about the 

 seven anterior abdominal segments, each of which bears a spiracle. Behind the seventh, 

 however, there is a somewhat confused mass of segments, bearing a single spiracle on its 

 anterior part; this spiracle is usually larger than those upon the other abdominal 

 segments (see woodcut 4J. This confused mass is sometimes described as a single 

 segment and sometimes as two. Careful comparison with the pupa proves that it 

 is certainly made uj) of three segments. 



Woodcut 3, X -', represents the posterior part of the larva of Gonoptera libatrix, as seen 

 from the left side. The separation of the anterior, spiracle-bearing part of tlie mass of 

 segments behind the seventh abdominal (vii) is extremely distinct in this larva, clearly 

 forming the eighth abdominal segment (viii). Behind this there is a small ninth abdo- 

 minal (ix), which is clearly separated off dorsally, although less distinct ventrally. 



The pairs of dorsal bristles shown upon the eighth and ninth abdominal segments in 

 woodcut 3 are obviously homologous with those on the anterior abdominal segments. In 

 woodcut 4, X 0, the same parts are seen from the right side in the larva of a Pyrale, 

 Ephestia Kiilimiella ; the ninth abdominal is seen to be even more distinct than in wood- 

 cut 3, and to be clearly defined ventrally. In the larva of Aglla tau the independence 

 of the ninth abdominal (and the tenth also) is strongly confirmed by the resemblance of 

 the pair of dorsal tubercles to those upon the anterior dorsal segments *. The accurate 

 representation of these parts in many South- American larvse should be studied in the 

 plates of W. Miiller's ' Siidamerikanische Nymphalidenraupen ' (Fischer, Jena, 18SG). 



In the pupa, this ninth abdominal segment, although small, is as distinct as any of the 

 others; see, for instance, woodcut 5, x7, which represents the last three segments of a 

 female pupa of Pieris brassicce, seen from the left side : the same parts of the same pupa 

 are seen from the dorsal aspect in woodcut 6, x7. The distinct independence of the ninth 

 abdominal is shown with equal clearness in most of the figures on Plates XX. & XXI. 



The part behind this segment in the larva forms a tenth abdominal. This segment 

 is separated into a dorsal portion (x' in woodcuts 3 & 4), of which the posterior 

 and lower part forms the anal flap, and a ventral portion (x), of which the anal claspers 

 form the posterior and lower part ; between the latter is the anus. 



2. The Relation of the Terminal Abdominal Segments of the Fupa to those of the Larva. — 

 The essential structure of the terminal parts of the vast majority of pupae, as seen from 

 the left side, is diagrammatically shown in woodcut 7. The identification of the various 

 parts with those similarly numbered in the larva (woodcuts 3 & 4) is sufficiently obvious. 

 The spiracle on the eighth abdominal is rudimentary in the pupa, although of exceptional 



* See Poulton in Trans. Ent. Soc. London, 1888, p. 561 ; woodcut 4 is copied from plate xvii. fig. 9 accompanying 

 that paper. 



SECOND SERIES. — ZOOLOGY, VOL. V. 30 



