THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 545 



yellow chromatophores are expanded a dull yellowish-brown 

 ground colour results (Plate 23). 



The vertical stripes are due to the development of black 

 and red chromatophores alono-six somewhat irregular bands, 

 beginning just in front of the dorsal fin and ending at the 

 base of the tail fin. These bands of chromatophores are of 

 considerable interest. They occur in their most marked form 

 in the superb cross-striping of coral reef Labroids and other 

 families, but they also appear under stimulation as a series of 

 evanescent banded markings on the skin of unstriped fish. 

 The common Crenilabrus rupestris shows this vei-y well. 

 When at rest it is of a nearly uniform brown or dull reddish 

 colour, but on being handled or when transferred to con- 

 trasting surroundings the body is seen to be overspread in a 

 Avave-like manner by bars of a deeper colour, which may con- 

 tinue to come and go in blushes. Again, Holt has recorded 

 the appearance and disappearance of dark transverse bars in 

 the common ballan wrasse (Labrus raaculatus). In this case 

 the fish had exactly the same property of expanding and con- 

 tracting the metameric tracts of chromatophores without 

 altering the body colouring. Sometimes, indeed, the bands 

 disappeared almost entirely and the fish became of a uniform 

 green colour. Crenilabrus rupestris has, at least in its 

 younger stages, the same property. It may, and usually does, 

 exhibit a banded appearance, but the bands may be extin- 

 guished and the body assume an almost uniform green tinge. ^ 



The presence of these bars of colour is b^ no means wholly 

 dependent on the nature of the surroundings. In Creni- 

 labrus melops they tend to appear under conditions that 

 favour expansion of pigments, but they also appear instantly 

 if a fish is transferred from white to dark vessels. Tactile 

 stimuli are especially effective in bringing about alternate 

 flushing and pallor along these tracts. It is clear that they 

 are more or less raetamerically arranged tracts along which 



* Since this was written the observations of Townsend (1909) and of 

 Tate Regan (1909) have revealed an unexpectedly wide range of rapid 

 colour-changes in tropical fishes. 



