Phenomena connected with Insect Metamorphosis. 161 



roughnesses on the body of the larva itself; or they resemble 

 the patches of lichens and crevices in the bark of the larger 

 branches, as also do the thick bunches of hair on certain 

 larvae. In those brown larvae again, which hide about the 

 roots of plants, the markings harmonise with the tints of the 

 earth and dead leaves. At first sight it is not quite so easy 

 to account for the markings found on larvae having a green 

 ground colour, but this difficulty disappears on a little con- 

 sideration. Those marks usually take the form of longi- 

 tudinal or diagonal lines. ISTow suppose we have a green 

 larva, so long as it is small, it is sufficiently well concealed 

 by the resemblance of its colouring to that of the leaves of 

 its food plant, but when it increases in size, there arises the 

 necessity, if it is to escape notice, of some means being pro- 

 vided whereby the attention may be drawn from the outline 

 of the larva, and it is in these markings that we find this 

 means: first, with regard to longitudinal lines, these are 

 most common in caterpillars feeding on narrow-leaved plants, 

 and when they occur in those feeding on large leaves it is 

 only in medium-sized individuals ; the reason for this being 

 that these lines mimic the lines found on grasses, etc., and 

 the streaks of light occurring in masses of foliage. In a 

 small larva such lines would be comparatively useless, and it 

 is for this reason that they seldom make their appearance 

 till after the third moult. Again, in broad -leaved plants 

 there are of course no lines on the leaves corresponding to 

 these longitudinal lines, nor do we commonly find long narrow 

 streaks of light let in among such foliage ; so that in larvae 

 feeding on those leaves, it is obviously most beneficial that such 

 lines should only occur on those of a medium size, or, when 

 on a large larva, should be broken and not continuous. On 

 the other hand, in grass-feeding caterpillars, as the Sati/ridce, 

 we find this arrangement strongly developed, as also among 

 many pine-feeders, in some of which, as Fidonia piniaria, the 

 green is striped with white; in others, as Anceryx, with 

 brown, thus resembling the glimpses of the brown twigs seen 

 among the green spikes. This arrangement, I may mention, 

 exists also in certain saw-fly larvae, while in DendroUnus 

 pini, in addition to this, there are tufts of hair closely imitat- 



