98 THE INFLUENCE OF INANIMATE SURROUNDINGS. 



a few contradictory statements. He says that when a Gobiiis 

 Ruthensparri is placed on a red bottom the yellow chroma- 

 tophores shrink as well as the black ones, although the yellow 

 contract less strongly than the latter ; but, according to 

 Pouchet's explanation, the yellow chromatophores should hardly 

 or never contract under a red light, since it is incapable of 

 affecting even the red chromatophores. This indicates that 

 there is still much to be done in this enquiry ; and it is to be 

 hoped that naturalists who take an interest in the subject and 

 are in a position to make independent investigations will not 

 suppose that it is exhausted even after the interesting and 

 extended experiments of Lister and Pouchet. 



We must also guard against the idea that another question 

 which is connected with this has been in any way answered : 

 namely, that as to the first formation of the pigment in the 

 chromatophores a question which is often, but erroneously, 

 regarded as identical with the other : How a particular mode of 

 coloration, or rather of distribution of the pigments, is to be 

 accounted for. This has, in fact, been fully explained by Lister 

 and Pouchet in the case of chromatic function, but it is clear 

 that the other question is not touched by it ; for chromato- 

 phores, i.e. dermal cells characterised by a rapid and peculiar 

 contractility, must have existed before the contractions occa- 

 sioned by the light reflected from surrounding objects could 

 result in a useful function. The permanence and even the 

 further development of the chi-omatic function in such animals 

 as most reqiiired its protective effects is of course easily ex- 

 plained by the principles of the Darwinian theory by natural 

 selection in the struggle for existence ; but its first occurrence 

 depends exclusively on the pre-existence of pigment in highly 

 contractile cells. 



The contractile power of the chromatophores, however, offers 

 no special difficulty, as has already been observed, since we 

 know that protoplasmic cells, devoid of an enclosing membrane 

 like those of the chromatophores, are universally endowed with 

 this property ; any such cell, being a cell of the connective 

 tissue of the cutis, might become a chromatophore, if pigment- 

 grannies were deposited in its protoplasm. Thus the only 



