if)ii.] Gravely & Mauwk : Trec-hauntins, Neuroptera. 103 



removed from the bark it will remain absolutely motionless for a 

 time, feigning death. 



When thus at rest on the bark the legs are entirely hidden 

 beneath the body, and the long powerful mandibles are drawn 

 back and held so widely open as to lie beneath the sides of the head 

 and the lateral processes of the thorax. If, when the larva is 

 hungry, a fly happens to touch it on any part of the head the 

 mandibles are closed instantaneously and the fly captured be- 

 tween them. No suggestion of discrimination as to the qualities 

 of the fl}^ was ever observed in the process. After this the fl}^ is 

 shifted along towards the distal end of the mandibles, and if it is 

 dropped before reaching its destination no effort is made to recover 

 it. If the end is reached in safety the mandibles are thrust into the 

 body, always between two segments — apparently they are too 

 blunt to pierce any harder part of the integument. The sucking 

 of the juices of the fly along the canal on the under side of each 

 mandible is then commenced at once and the piston-like motion 

 in these canals, b}'' which suction is effected, may be seen under 

 a hand-lens. From time to time one or other of the mandibles 

 is withdrawn in order to commence sucking in another place, the 

 fly being held aloft and quite clear of the bark throughout the 

 whole time of feeding. Flies continued to move for a long time 

 after they were caught ; they did not appear to be poisoned as 

 did those caught by Mr. S. Green's species in Ceylon (Westwood, 

 1888, p. 8). A fly is finished in from half an hour to an hour. 

 The mandibles only — never the legs — are used in manipulating it. 



Cocoons and Pupae. 



The Pseudoptynx larva constructed its cocoon at the surface 

 of the loose dry soil provided, by fastening together pieces of earth 

 with tough silk (fig. 15). 



The larvae of Myrmeleon contractus spun cocoons in crevices of 

 the bark on whichthey were living, although all other My rmeleonids 

 of which the cocoon isknown appear to spin in soil. Having found 

 a suitable crevice the larva sits in it with the head erect and jaws 

 projecting upwards, and proceeds to spin round the edge with silk 

 extruded from a retractile spinneret at the posterior end of the 

 abdomen, the abdomen being moved to and fro throughout the 

 process. The edges of the cocoon become broader and broader, 

 being carefully covered with dust as they are elaborated, and the 

 aperture in the middle becomes narrower till finally the jaws are 

 withdrawn and the cocoon or at least its outer covering com- 

 pleted. If a cocoon be opened it is found to consist of two layers 

 of silk, the inner one being softer and more loosely spun together 

 than the outer. When the mature insect emerges the pupal skin is 

 left projecting from the aperture made in the cocoon (see figs. 9 

 and 10). Presumably the pupa eats its way through the silk with 

 its peculiar jaws (fig. 12) as suggested by Westwood in the case 



