22 MR. A. HAMMOND ON THE 



anterior spiracle of the larva immediately under the integument, 

 and partially surrounding the anterior termination of the main 

 trachea. I think that in this case similarity of development is 

 a strong argument in favour of similarity of homological relation- 

 ship ; and again it may be asked, if they be not the homologues of 

 the wings, how are we to regard them ? To look upon them as 

 abnormal productions would, I submit, be contrary to the whole 

 spirit of philosophical inquiry ; and what other opinion we can 

 form I know not. If, then, they be the proper dorsal appendages 

 of the prothorax of the pupa, then the imaginal structures found 

 immediately underneath them must in all probability correspond, 

 and be prothoracic too. But these structures are the humeri to 

 which I have had occasion to refer. Therefore, with Burmeister, 

 I must look upon these parts as prothoracic*, and consider them 

 as the homologues of the posterior angles of the collar of the Hy- 

 menoptera, the homologous parts in both orders being followed 

 immediately by the spiracle. 



But it is not only in the prothorax that the observation of de- 

 velopmental change will afford a clue to the division of the seg- 

 ments. In the pupa of the Crane-fly the dorsal surfaces of the 

 meso- and metathorax are sufficiently and distinctly marked, the 

 former being as conspicuous for its extent as the latter for its 

 contracted dimensions ; and, strange to say, their dorsal appen- 

 dages are not yet recognizable as a pair of wings and a pair of 

 halteres, but as two pairs of undoubted wing-cases similar to each 

 other in every respect but that of size. It is only when we sepa- 

 rate the latter pair and examine them carefully with a lens that 

 we can persuade ourselves that the nascent organs within them are 

 not really wings, but the familiar halteres (see PI. II. figs. 3 & 4). 

 They are, so far as I judge, unquestionably modified and abortive 

 posterior wings, appendages of a metathoracic segment, however, 

 reduced, and by no means abdominal, as was supposed by 

 Latreille. 



Again, on carefully removing the integument from the dorsal 

 surface of the Crane-fly pupa over the posterior portion of the 

 mesothorax, in front of its junction with the metathorax I 

 disclosed the plate marked ps in the drawing of the imago 

 (PI. II. fig. 2), which I must therefore regard as mesothoracic. 

 This plate is nearly horizontal in the Crane-fly ; but a compa- 



* Antea, p. 12. 



