110 LEAF-MINING INSECTS 



half of the cocoon it cannot retreat any further and instead it 

 contracts the body more and more. . . . But when the new 

 portion of the cocoon is so advanced that the body can undergo 

 no more contraction for applying the head to its edge then it bends 

 the fore part of the body considerably. It turns the head back 

 over the body and so is still in a position to apply the spinneret 

 to the border of the cocoon. . . . When the distance between 

 the edges of the two halves of the cocoon becomes very small 

 there is only room for the head between them, and at last this 

 distance diminishes to such an extent that there is no longer room 

 for the head to move without disarranging the tissue of the edges. 

 The caterpillar was then obliged to close this opening in another 

 way. ... It changed entirely its manner of working. It drew 

 its head into the cocoon and then stretched threads of silk between 

 the two edges parellel with the long axis of the cocoon. It added 

 other threads to these and in this manner it soon joined together 

 the edges of the two portions of the cocoon. In all the finished 

 and complete cocoons it is easy to recognize the place where the 

 two parts have been joined together where there is always a little 

 area across the cocoon in which the ridges or cords of the two 

 halves do not exactly correspond, where they do not fall quite in 

 line. . . . 



The construction of the tissue itself is noteworthy. When the 

 cocoon is entirely finished it then appears only as a closely woven 

 tissue, marked with longitudinal ridges or cords. . . . But let 

 one examine the cocoon with the aid of a microscope when it is 

 only begun or when only half is done or while it is still not entirely 

 closed, and he will see that the threads attached to one of the ribs 

 pass to the next one and so on. There are two sets of these threads 

 between each pair of ribs and these extend obliquely, but the 

 threads of one set are arranged at an angle to the threads of the 

 other set so that those of the first are crossed by those of the other 

 and thus they form meshes like those of fishing nets. They ad- 

 here together in the places where they cross one another 



That the threads are not stretched between the ribs as are the 

 strings on a violin or a harpsicord, but that they are lax and that 

 they have a downward curve and thus together form a little groove 

 between the ribs The half cocoon at which the cater- 



