285 



oxygen. The larvae of Tanytarsus form a characteristic case attached 

 to stems of grass or other objects in the water. 



Transformation to the pupa generally takes place within the tun- 

 nel, the sides of which consist of the saliva of the larva, which seems 

 to harden on contact with the water, and in which there is generally no 

 trace whatever of threads. During the last larval instar the develop- 

 ment of the imaginal disc is very rapid, and in a series of larva} of 

 any species taken at the same time it is not unusual to find specimens 

 which represent an almost continuous series of the changes which take 

 place. The most striking thing about the transformation is the devel- 

 opment of the imaginal head. In the earlier stages of its formation the 

 head with all its parts lies within the larval head; but gradually, as 

 growth proceeds, it is so withdrawn that the compound eyes lie outside 

 of the larval head and within the larval prothorax. The imaginal head 

 is generally conspicuously larger than that of the larva, which in a 

 measure explains why the complete transformation does not take place 

 within the latter. Miall and Hammond have dealt at considerable 

 length with the transformation here referred to.* It is exceptional to 

 find a species in which the head of the imago remains entirely within 

 that of the larva for the greater part of its period of formation, though 

 some cases of this kind are recorded. f 



I have not met with any species — even in Tanypincu, in which the 

 larval head is larger than in other groups — in which the head remains 

 long within the larval head after transformation begins, but I have 

 found one specimen which in some unaccountable way had failed to 

 withdraw the head in time, and, in consequence, the head with its mem- 

 bers was tightly compressed within the cavity of the larval head, the 

 neck being much elongated. The specimen was in alcohol, so that 

 it w^as impossible to say whether or not it could have successfully 

 emerged. 



At the time of the emergence of the pupa the Inrval skin splits 

 longitudinally on the dorsum of the thoracic segment, and generally 

 along the middle of the dorsal surface of the head. Miall and Ham- 

 mond state that at this time the head splits along the central suture,t 

 which feature I have observed also, but in some cases both the dorsal 

 and ventral surfaces are ruptured. I have not sufficient material to de- 

 cide whether there is any distinguishing character in the rupturing of 

 the head in the different species or genera. 



*The Harlequin Fly, pp. 118-137. 1900. 

 + Miall and ITannnond, 1. c, p. 135. 

 tLoc. cit., p. 27. 



