THE COLORATION OF REEF FISHES 593 



possessor by its color and pattern under natural conditions have 

 not been undertaken. Young's ('16) feeding experiments, which 

 among those coming to my attention, a few imperfectly reported 

 by Poulton and Sanders ('98) excepted, alone possess positive 

 value, suggest, however, that the latter has directed the develop- 

 ment in question. 



It is not clear that all advocates of the hypothesis of direct 

 action would maintain its sufficiency to account for such de- 

 tailed agreement in color as appears to exist, between organism 

 and environment in some cases. Allen ('77, '07), referring 

 chiefly to North American birds, writes: 



The southward increase of color correlates with an increase of at- 

 mospheric humidity and temperature, and consequently with the 

 protective influences of luxuriant arboreal vegetation and clouds; 

 and, conversely, loss of color accompanies excessive aridity, a scanty 

 vegetation, and an almost cloudless sky, the conditions, in short, of 

 all others the most powerfully effective in the blanching of color; and 

 again, the somber, dusky tints of the northwest coats accompany 

 the most humid conditions of climate and the conditions generally 

 most favourable for the protection or preservation of color. 



This excerpt seems to imply appreciation of no such agreement 

 in color between birds and their environment as exists in general 

 among the reef fishes; yet it may, perhaps, be fairly inferred 

 from Chapman's ('94) observations upon the birds of Trinidad 

 that the truth regarding them is the same. If the inference is 

 correct, it does not appear that Allen's hypothesis can explain 

 existing conditions, for it involves nothing more on the one 

 hand than reduction of color through blanching, and the pres- 

 ervation, upon the other, of any hues, congruous or otherwise, 

 which may be developed in the process of metabolism. Pack- 

 ard's ('04) conception of the relation between organism and 

 environment, upon the contrary, postulated more definite con- 

 trol of the creature's chromogenic function, since he held that 

 alleged cases of Mullerian mimicry might be explained by con- 

 vergence due to the action of siinilar physical or climatic causes. 



All the facts which the present research has yielded conform 

 to the terms of Thayer's hypothesis of concealing coloration. 

 How, then, has it happened that twenty years has failed to bring 



