THE COLORATION OF REEF FISHES 557 



664), led to careful study of the habits of the five species of red 

 fishes which occur in the shallow waters of the inner reefs at 

 Tortugas. 



The conclusion that these animals belong to a well defined 

 ecological group was soon reached, and is supported by the 

 following facts. 



The fishes are very rarely seen in the open by day. Pria- 

 canthus cruentatus has not been seen fully exposed, of its own 

 initiative, except at or after tA\'ilight. One specimen of Amia 

 sellicauda was observed a few inches from cover, and the squirrel 

 fishes, Holocentrus ascensionis, siccifer and tortugae, are little 

 less retiring in habit, for weeks may pass without one being 

 noted outside the shelter of the coral stacks,^ although the 

 observer may be on the alert to catch them. 



The five species may be secured in the daytime b^i blasting 

 with dynamite at almost any station among the stacks in whose 

 crevices they lie hidden. As many as six indi\dduals, represent- 

 ing four of them, have been taken at once, and complete failure 

 rarely followed when using dynamite among the heads. In 

 attempting to visualize what this means regarding the com- 

 parative abundance of hidden and exposed specimens one should 

 bear in mind that the explosive, in the quantity used, is effective 

 up to a distance of only 4 or 5 feet from the point of discharge. 

 But the fishes taken are not uniformly distributed throughout 

 its 'sphere of influence,' which lies largely outside the heads 

 among which they lurk. This is apparent from the fact that 

 where the work was done the water averaged little more than 

 8 feet in depth. In addition, it was always necessary to place 

 the shot within a yard of the outer face of the stack, which 

 in water of the depth mentioned, commonly rises nearly verti- 

 cally to a height of 5 or 6 feet. If this precaution were not taken, 

 the fragments of coral, failing to be thrown out into the open, 

 would fall in a heap from which the specimens could not be re- 



" The coral stacks are masses of heads rising nearly vertically from the bottom 

 in water from eight to fifteen feet deep. They are, in respect to the shelter they 

 afford to fishes of sm^ll and medium size, upon a level with a pile of boulders 

 loosely cemented together. 



