MOEPHOLOGY OF THE LEPIDOPTEKOUS PEPA. 197 



the tenth, and separating the rostral segment (x' in woodcut 7) as the eleventh from 

 any contact with the ninth. Now, however, I think it is far more probahle that the 

 line of separation between the two parts of the tenth abdominal in woodcut 7 merely 

 corresponds to the posterior part of the chink beneath the larval anal flap. The con- 

 striction which in certain pupae encircles the base of the terminal spine would then cease 

 to have any morphological significance ; and this is also rendered probable by the fact 

 that closely allied pupae are altogether without it. The existence of a distinct line 

 separating the supposed anal segment into a ventral and dorsal part in the condition 

 represented in woodcut 8, and the indication of such a division in the more pronounced 

 condition represented iu woodcut 9, are also irreconcilable with the view that the terminal 

 spine of these woodcuts represents the whole of the dorsal part of the tenth abdominal 

 (x' in woodcut 7). 



We may therefore conclude that both larva and pupa possess ten abdominal segments ; 

 and eveu if this conclusion may require subsequent modification, the segmental 

 relations of the external reproductive organs will be unaffected ; for these structures 

 only come into relation with the eighth, ninth, and the ventral (anal) part of the tenth 

 abdominal segments. 



Part III. — The External Generative Organs. 



1. Introductory and Historical. — It is very remarkable that these organs have not been 

 universally recognized, considering that thousands of pupae are examined every year by 

 entomologists and dealers, while the organs are easily seen in large species by the naked 

 eye. There are other well-known tests of the sex of pupae, depending on the difference 

 between male and female antenna?, and upon the greater size of the abdomen in the 

 female. But the former only applies to certain species (in which such a difference 

 exists), and in these the antennae of the pupae are far more alike in the two sexes than 

 those of the imagines of the same species (as will be shown in a future paper). The latter 

 test is even more precarious. I therefore think that a study of the sexual differences 

 on Plates XX. & XXI. will be of practical assistance in many departments of the 

 subject. I have myself found the knowledge invaluable for many years, in the manage- 

 ment of some experiments with larvae in which it was necessary to ascertain and allow 

 for hereditary tendencies. The pupae of moths can be arranged according to their sexes 

 far more quickly than the imagines of the same species, except when there is some 

 obvious secondary sexual character, such as a difference in the colouring of the wings, 

 &c. The organs are especially easy to distinguish, because pupae are either entirely hair- 

 less, and generally smooth in the region of these structures, or possess minute or scanty 

 hairs which do not cause any obscurity. 



In the first edition, now out of print, of ' Forms of Animal Life' (1870) by the late 

 Professor Rolleston, these words occur on p. 76, in a description of the pupa of Acherontia 

 atropos : — " The ninth abdominal ring is marked by a depression on either side of the 

 middle ventral line, the lines limiting which extend into the interspace between it and 



30* 



