CH. III.] INSECTS WHICH FORM COCOONS. 63 



they constantly work. He dug up one of these co- 

 coons when it was nearly completed, and broke off 

 about one third of its extent at one end, as in fig. A, 

 from which it will be perceived that the thickness 

 of the structure is considerable. The insect imme- 

 diately commenced the reparation of its cocoon, 

 which it completely effected in the course of four 

 hours, by means of bits of earth which had been 

 placed within its reach. It commenced by protru- 

 ding nearly the whole of its body out of the breach 

 (fig. B) and seizing a piece of earth with its strong 

 jaws, which it bore into its damaged cocoon ; this 

 employment it continued for some time, selecting 

 with much care those bits of earth which were fit- 

 test for its use ; during this time, therefore, it was 

 employed only in collecting materials, since it was 

 rarely that it affixed any of these grains to the walls 

 of its cell unless it met with one which happened 

 exactly to fit any part of the damaged breach. It 

 now, however, commenced spinning a slight band 

 of silken network round the aperture, and then at- 

 tached bits of earth thereto by means of silken 

 cords, which it spun for that purpose during its la- 

 bour ; thus by degrees the size of the damaged part 

 became contracted. When the aperture had be- 

 come nearly closed (fig. C), the caterpillar, in order 

 to close it entirely, was observed to spin arched 

 lines of silk from one side to the other of the aper- 

 ture, crossing each other at different angles, like a 

 coarse kind of network, into which it fastened 

 grains of earth, so as to render its external surface 

 similar to the sound part ; it however did not con- 

 tent itself until it had also strengthened the inside 

 with smaller grains of earth, so that when Reaumur 

 opened the cocoon again with a penknife, the new 

 part was found to be equally thick and compact with 

 the other part which had not been damaged. The 

 internal operations of these caterpillars are not ob- 

 servable, because they do not commence the silken 



