THE COLORATION OF REEF FISHES 535 



The diversity of coloration occurring among reef fishes led 

 to the testing of this assumption by its application to them, but 

 in such an undertaking no advantage is to be won thiough 

 multiplying sporadic instances of 'protective' or Svarning' color- 

 ation which may attract one's attention upon the reefs. It is 

 necessary rather, if any should appear to exist, to determine the 

 system or law of distribution of color among the fishes of a given 

 region, in the hope of discovering relations of general validity, 

 facts which, regardless of preconceived opinion, shall make the 

 same appeal to all observers to whose attention they may come. 

 It is not clear that there is more than one way to approach such 

 a problem: the foundation of all later researches, pursued by 

 whatever method, must be laid by long continued study of li\'ing 

 speciniiens under absolutely normal conditions. 



It may be an incorrect conclusion based upon ignorance of 

 other animals, but tropical reef-fishes seem superior material 

 upon which to carry out such a preliminary study as the present. 

 In the richness of their coloring they vie with birds and butterflies. 

 Many species may be observed in one circumscribed region. Of 

 these a fair proportion are common. The great majority of 

 them are not very shy and are little disturbed by observation 

 from a boat, and perhaps as little, or even less, by one w^ading 

 among them waist-deep in water or going down among them at 

 greater depths with diving equipment. They possess, as a class, 

 large power of color-change, which, correctly interpreted, is of 

 the highest significance, for it provides an unmistakable clue to 

 much that has remained obscure. They may be studied in a 

 perfectly natural environment, unaffected by human activities 

 as most terrestrial habitats are. Finally, their food is such that 

 its nature may usually be determined without difficulty from 

 analyses of stomach contents, and conclusions drawn from 

 observation of the creatures' movements be checked readily 

 against information of this sort and validated or disproven. 



Regarding methods of observation little need be noted in detail; 

 such information as is necessary or of interest will appear in its 

 appropriate place with the statement of results obtained. The 

 fishes have been viewed from every angle above and below, by 



