290 Mr. 0. H. Latter on the secretion of 



examination reveals a far more efiicient apparatus. I 

 placed one individual in absolute alcohol the instant it 

 emerged from the cocoon, and so secured it with the 

 "shield" still on. Examination of this specimen showed 

 two small sharp points (figs. 1 & 2, a a) projecting in 

 front of, and just ventral to, the pupal labrum from the 

 labrum of the imago. These points are entirely con- 

 cealed by the abundant fiufiy scales on the head 'of the 

 fully developed imago. If the head is "plucked," they 

 are readily seen, even with the naked eye (see fig. ij. 

 The points are about 0.5 mm. in length, and about 

 0*75 mm. apart, and project forwards and downwards : 

 their outer margins are thickened, and the two together 

 are carried upon a slightly thickened elevation of the 

 imago's head. On each side of this elevation is an 

 elongate pit (fig. 2, hh), open towards its ventral aspect, 

 and running up dorsally towards the vertex of the head 

 to end about a third of the way up in a socket. The 

 total length from the socket to apex of points is about 

 1-2 mm. pn the inner surface of the pupal "shield," 

 i. e., the side which is in contact with the underlying 

 imago, is a pair of recurved hooks (fig. 4, hit), which 

 fit accurately into the groove and socket of each side — the 

 grooves serving as guiding lines towards the sockets. It 

 is thus impossible for the " shield" to be removed by any 

 upward pressure, for all pressure in that direction merely 

 serves to lock the "shield" more firmly to the head of 

 the imago below. On the other hand, downward pressure 

 causes the hooks to come out of the sockets, to slide 

 along the grooves, and the shield may be set free. 

 These appearances w^arrant the conclusion that the 

 imago attacks the softened inner surface of the cocoon 

 with the above-mentioned pair of pointed processes, and 

 that the "shield" serves to protect the eyes and head of 

 the imago from injury which might otherwise result 

 from friction against the walls of the cocoon. Probably 

 each stroke is made in a downward and forward direction 

 by the head of the animal, and probably fresh applications 

 of the potassium hydroxide are made as each softened 

 layer is in turn removed in order to soften the succeeding 

 layers. I believe the potassium liydro.N,ido is produced 

 from the mouth — at any rate I could discover no other 

 aperture of any kind : if this surmise be correct, it seems 

 probable that the muscular efforts necessary for scraping 



