606 Dr. D. Keilin on 



abdominal segment is prolonged beyond the anus so as to 

 form a respiratory siphon similar to, but much shorter than, 

 that of larvEe A. The larva is amphipneustic ; the pro- 

 thoracic spiracles (fig. 5) protrude slightly externally and 

 show traces of the spiracular papillae, Mhose number I could 

 not determine. As in Larvae A, the respiratory siplion is 

 bifurcated at its extremity, and, in consequence, the post- 

 abdominal spiracles are separated by tlie groove of the 

 bifurcation (fig. 7). The hucco-jiharynyeul armature of this 

 specimen resembles that of Larvse A, but the sclerites are 

 relatively more slender (fig. 8). 



The Systematic Position of Larva A and B. 



As already mentioned, these larvse had been associated by 

 Mr. F. W. Edwards with the Eristalinse (Syrphidse). He 

 supplied, moreover, the following information (see Baylis, 

 p. 379) : — '' Dipterous larvae from Cardiosoma hirtipes. — 

 These larvaj are evidently Syrphidse, and apparently belong 

 to the subfamily Eristalinse ; they differ from Eristalis iu 

 the more elongate form and the lack of any obvious 

 separation into 'body' and 'tail.'' 



" Larvae from Gecarcoidea lalandii. — These are also Syr- 

 phidae, but in the present state of our knowledge it is 

 impossible to assign them definitely to any subfamily. 

 They a|)pear to lack the extensile ' tail ' of Eristalis^ 



Tlie foregoing remarks show that Edwards was guided in 

 his identifications solely by the general conformation of the 

 body, and especially by the presence of a well-developed 

 posiabdomiual respiratory siphon. As a matter of fact, the 

 resemblance between larvae A and an Eristaline larva is 

 most remarkable. It must be remembered, however, that 

 the possession of a postabdominal siphon is merely evidence 

 of the fact that the larva inhabits a fluid or semi-fluid 

 medium, and the greatest caution must be exercised before 

 making use of it as a character for the purpose of systematic 

 classification. A well-develoned postabdominal respiratory 

 siphon is known to exist iu many totally diff'erent groups of 

 D.pterous larvae, e.g., Ptychopteridte, Psychodidse, Culicidae, 

 Stratiomyidae, Phoridae, Eristalinae (Syrphidae), Antho- 

 myidae, Ephydridse, Drosophilidte, and others. 



Being specially interested in the structure of Eristaliue 

 larvffi, I have had occasion to examine many examples of 

 this subfamily, and, from my knowledge of their morpho- 

 logy, I am forced to the conclusion that the larvae under 

 consideration have no affinity whatever with the Eristalinae, 



