THE PUPAL STAGE OF CULEX. 55 



allow the imago to escape. A pair of branched seta? arise from the 

 dorsal region of the hinder part of the thorax. 



The respiratory siphons (at, Fig. 1) are nearly cylindrical, nar- 

 rowed at their bases and curved forwards to be attached by 

 flexible membranes to slight prominences on the sides of the pro- 

 thorax. Above they are obliquely truncate and open, and the 

 margin is slightly notched on the inner side. The outer surface is 

 marked so as to resemble imbricated scales, each with a minute spine 

 at its apex. The cavity of the siphon communicates directly with 

 that of a tracheal trunk at its base. Palmen (6) says that after a 

 " close investigation " he has found that there is no opening. The 

 tone of assurance in which he contradicts all previous observers led 

 me to put the question to the test. I removed the side wall of the 

 thorax, with some of the underlying muscles and tracheae, from a 

 specimen preserved in alcohol. I drew out the alcohol from the 

 cavity of the siphon by means of blotting-paper, and then touched 

 the tip with a minute drop of glycerin. I watched the effect under 

 the microscope, and saw the glycerin force its way into the siphon, 

 driving the air before it into the tracheae. Palmen, moreover, says 

 the organs are gills ! Each is a thick chitinous tube, the cavity 

 guarded by numerous hooked spines, the walls consisting of hardly 

 anything but the chitinous cuticle, the epidermis (" hypodermis ") 

 between its two layers being barely recognisable on account of its 

 thinness. The "tracheal gills" on which Palmen lays much stress 

 have absolutely no existence. 



The icings of the pupa, that is the organs within which the wings 

 of the imago are developing, are a pair of oblong plates about 2| mm. 

 in length. They are closely applied to the sides of the hinder part 

 of the thorax, and directed downwards and backwards. They are 

 immoveable. 



The halter es are a pair of elongated triangular plates lying along 

 the dorsal and hinder border of the wings. 



All these three pairs of dorsal appendages arise within the larva 

 in the same way, and their bases or points of attachment all lie in 

 the same horizontal plane. Each is at first (in the larva) a fold of 

 the epidermis; each acquires a cuticular covering (like all other 

 parts of the body), and the first pair become rolled up to form 

 tubes, the respiratory siphons, while the other two remain flat plates. 



