ON SOME LEPIDOPTEEOUS LARV^. 91 



fifth body-segments are very black patches, which, when 

 the insect is at rest, are quite concealed by a flap of skin. 

 But when the terrifying attitude is assumed, these flaps 

 fall back and reveal the black patches. Hermann Miiller 

 has suggested that these marks serve to imitate the appear- 

 ance of ichneumon stings, so that the ichneumon fly may 

 conclude that the caterpillar has already been attacked and 

 tenanted by one of its race. Prof. Poulton found that these 

 caterpillars were not distasteful, but that they were eaten 

 both by a marmoset and a lizard with which he experi- 

 mented, though when in their defensive attitude they were 

 only approached by both animals after long examination and 

 with much caution. 



I next call attention to some specimens of larvse of the 

 swallow-tailed butterflies. They are very beautiful, and are 

 remarkable for the peculiar Y-shaped process which they can 

 protrude from behind the head, and which is no doubt an 

 intimidation to their foes. 



I have recently referred to cannibalism : for the credit 

 of the larval community, I am pleased to testify that this 

 only occurs in a few species, and close confinement in our 

 breeding- cages may sometimes bring it about when larvae 

 in freedom in their natural conditions would not be guilty 

 of it. There are one or two species, however, the cater- 

 pillars of which are inveterate cannibals. Amongst the 

 worst of them may be mentioned Cosmia trapezina ; this, if 

 accidentally introduced into the cage, will soon slaughter 

 and devour all its companions. From the way in which it 

 follows up and attacks them, it seems probable that this is 

 not a sudden impulse caused by dry food or confinement in 

 close quarters, but is most likely a characteristic of the 

 larva in its wild state. It devours not only larvae of 

 different species from itself, but its own brothers and sisters, 



