v Shifts for a Living 53 



Serranus ; and in not a few crustaceans. The researches 

 of Briicke, Lister, and Pouchet have thrown much light on 

 the subject. Thus we know that the colour of surround- 

 ings affects the animals through the eyes, for blind plaice, 

 trout, and frogs do not change their tint. The nervous 

 thrill passes from eye to brain, and thence extends, not down 

 the main path of impulse the spinal cord but down the 

 sympathetic chain. If this be cut, the colour-change does 

 not take place. The sympathetic system is connected with 

 nerves passing from the spinal cord to the skin, and it is 

 along these that the impulse is further transmitted. The 

 result is the contraction or expansion of the pigment in the 

 skin-cells. Though the path by which the nervous influence 

 passes from the eye to the skin is somewhat circuitous, "the 

 change is often very rapid. As the resulting resemblance 

 to surroundings is often precise, there can be no doubt that 

 the peculiarity sometimes profits its possessors. 



7. Special Protective Resemblance. The likeness 

 between animals and their surroundings is often very precise, 

 and includes form as well as colour. Thus some bright butter- 

 flies, e.g. Kallima, are conspicuous in flight, but become 

 precisely like brown withered leaves when they settle upon 

 a branch and expose the under sides of their raised wings ; 

 the leaf-insects (Phylliuin) have leaf-like wings and legs ; 

 the " walking-sticks " (Phasmid<z\ with legs thrown out at 

 all angles, resemble irregular twigs ; many caterpillars (of 

 Geometra moths especially) sit motionless on a branch, 

 supported in a strained attitude by a thin thread of silk, and 

 exactly resemble twigs ; others are like bark, moss, or 

 lichen. Among caterpillars protective resemblance is very 

 common, and Mr. Poulton associates its frequent occurrence 

 with the peculiarly defenceless condition of these young 

 animals. "The body is a tube which contains fluid under 

 pressure ; a slight wound entails great loss of blood, while 

 a moderate injury must prove fatal." " Hence larvae are so 

 coloured as to avoid detection or to warn of some unplea- 

 sant attribute, the object in both cases being the same to 

 leave the larva untouched, a touch being practically fatal.' 5 

 Among backboned animals we do not expect to find many 



