upon lepidopterous larva, dc. 297 



considerable distance by forcing it through the narrow 

 chink with its closely approximated lips, which con- 

 stitutes the mouth of the duct leading to the sac. Such 

 a formidable means of defence may readily have sup- 

 planted the more usual method of eversion, a method 

 which can only give rise to the discharge of vapour into 

 the air, instead of a well-directed stream of fluid, which, 

 if volatile, as it is in these larvae, of course produces 

 abundance of vapour. 



The appearance of the everted gland is indicated in 

 Plate X., fig. 7 (x 2 diameters), in which the head 

 and prothorax of a larva in the sensitive period are 

 shown from below. Of the four gland-tubes the anterior 

 pair are seen to be much larger and longer than the 

 posterior pair. The position of the structures corre- 

 sponds with that of the horizontal slit-like mouth of 

 the short duct leading to the sac, through which 

 eversion takes place, and which is shown upon the 

 red prothoracic margin of the head, below the true 

 mouth, in fig. 6. In an earlier paper (Trans. Ent. 

 Soc. Lond., 1885, Pt. II., August, p. 322) I have given 

 an account of the eversion and introversion of the pink 

 flagella of the larva of D. vinula. I have now studied 

 the subject more thoroughly, and can add further details. 

 In Plate X., fig. 8 (x8 diameters) the left conical 

 receptacle is seen from the left side, with its flagellum 

 completely everted. Immediately above the receptacle 

 the basal section of the flagellum is white or very 

 faintly pinkish, and through this part the rest of the 

 flagellum can be seen to pass during eversion and intro- 

 version. The withdrawn flagellum can also be seen with 

 a little care in the receptacle itself. It follows from 

 such a constitution that the summit of the receptacle 

 when the flagellum is introverted represents the rim of 

 a tubular depression, of which the lumen is bounded 

 by the morphologically outermost, — the cuticular layers 

 of the flagellum itself. In eversion the rim rises to 

 successively higher levels as each section of the outer 

 part of the flagellar cuticle passes over its edge, and 

 becomes truly outermost in position, and, last of all, the 

 apex is unfolded, and the rim then disappears. Con- 

 versely, in introversion the apex sinks, and a rim at 

 once appears, which also sinks until the summit of the 

 receptacle is reached. In fact, the structure affords 



TRANS. ENT. SOC. LOND. 1887. — PART III. (SEPT.) Y 



