332 INSECT TRANSFORMATIONS. 



width of the cells preventing them from stretching 

 the thread up to the top. To ascertain this he dis- 

 lodged several royal grubs about to spin their 

 cocoons, and introduced them into glass cells blown 

 of varying dimensions. ' They soon prepared to 

 work,' he says, c and commenced by stretching the 

 fore part of the body in a straight line, while the 

 other was bent in a curve, thus forming an arc of 

 which the sides of the cells afforded two points of 

 support. It next directed the head to such parts of 

 the cell as it could reach, and carpeted the surface 

 with a thick bed of silk. I remarked that the threads 

 were not carried from one side to another, which 

 would have been impracticable, for the larvae, being 

 obliged to support themselves, had to keep the pos- 

 terior rings curved; and the free and moveable part 

 of the body was not long enough to admit of the 

 mouth reaching the opposite sides. The first expe- 

 riments obviated the probability of any particular in- 

 stinct in the royal larvae, and proved that they spin 

 incomplete cocoons, because they are forced to do so 

 by the figure of their cells. But desirous of evidence 

 still more direct, I put them into cylindrical glass 

 cells, where I had the satisfaction of seeing them 

 spin complete cocoons in the same manner as the 

 larvae of workers. In fine, I put plebeian larvae into 

 very wide cells, and they left the cocoon open, as is 

 done by the royal larvcP. I also found that royal 

 larvae, when lodged in artificial cells, where they can 

 spin complete cocoons, undergo all their transforma- 

 tions equally well. Thus the necessity which nature 

 imposes on them of leaving the cocoon open, is not 

 on account of their increment; nor does it appear to 

 have any other object than that of exposing them to 

 the certainty of perishing by the wounds of their 

 natural enemy ; an observation truly new and siq-* 

 gular.'* 



* Huber on Bees, p. 133, 



