THE RELATION BETWEEN LIGHT AND PIGMENT-FOKMATION. 551 



was not perceptibly different from that of those kept on red 

 backgToiinds, nor from control specimens kept in clear vessels 

 uniformly illuminated. Larvie of Lepidogaster gonanii 

 and of L. bimaculatus were equally intractable. There 

 can be no doubt that long-continued experiments are neces- 

 sai-y owing to the slow pigmentary changes in these animals. 



(6) Summary. 

 (Crenilabrus melops.) 



Darkness (for five days) produces extreme contraction of 

 all chromatophores. 



Black and white backgrounds in white light (three weeks' 

 exposure) gave respectively the usual dark brown and the 

 light (green) colouring associated with this illumination. 



Brown-weed background acts like black. Green and red 

 algal backgrounds produce a greenish tint intermediate 

 between the effects of black and white backgrounds. 



Weed light-filters produce an entirely different effect from 

 the weed backgrounds (Table II). Daylight transmitted 

 through green weed induced brown coloi-ationand considerable 

 amount of new red pigment. Daylight transmitted through 

 red weed produced green coloration and yellow pigment. 

 Brown weed is too opaque for any differential effect to show 

 itself. 



These results, then, go to show that the action of algal 

 backgrounds is complicated by the impurity of the colours 

 transmitted or reflected from them. The nearer these approach 

 to the complexity of white light the more does the background 

 resemble a black or a white one, i.e. general contraction or 

 expansion results. The purer the colour or the more intense 

 the particular part of the spectrum, the greater is the develop- 

 ment of pigment of complementary colour. These results, 

 however, were too few to establish the relation, and I there- 

 fore undertook the following experiments on a more con- 

 venient subject, Hippolyte varians. 



