LEPIDOPTERA. 271 
jend of one of these leaves has, it appears, been rolled back towards 
ithe interior surface, in order to describe the first turn of a spiral, 
which is then covered by many other turns (Fig. 282). Some 

Fig, 282.—Oak leai rolled perpendicularly. Fig, 283.—Oak leat rolled sideways 
eaves are rolled towards their exterior surface, others are rolled 
sowards their interior surfaces, but in a totally different direction. 
he length or axis of the first roll is perpendicular to the principal 
‘ib and to the stalk of the leaf, the axis of the latter parallel to the 
same rib (Fig. 283). Work of this kind would not be very difficult 
i perform for those who had fingers ; but caterpillars have neither 
lingers nor anything equivalent to fingers. Moreover, to have 
tolled the leaves is only to have done half the work : they must be 
retained in a position from which their natural spring tends con- 
tantly to draw them. ‘The mechanism to which the caterpillars 
have recourse for this second part of their work is easily perceived. 
We see packets of threads attached by one end to the surface of 
he roll, and by the other to the flat surface of the leaf. They are 
;0 many bands, so many little cords which hold out against the 
pring of the leaf. There are sometimes more than from ten to 
iwelve of these bands arranged nearly in the self-same straight 
ine. Each band is a packet of threads of white silk, pressed one 
gainst the other, and yet we must remember all are separate.’’* 
Réaumur made the oak-leaf rollers work in his house. He has 
| * Mémoires pour servir a l’histoire des Insectes, tome ii., page 240 (Se Mémoire). 

