78 



feeders, and the other group depending largely on insects and crustaceans 

 and using mollusks sparely, and then only the small and thin-shelled 

 sorts. A similar inverse relation is seen between the large mouths and 

 the weak pharyngeals of many piscivorous fishes: Jietween the weak 

 pharyngeals and the muscular stomach of the gizzard shad, and between 

 the long gill-rakers and the rudimentary pharyngeals of the shovel fish. 

 8uch correlations are often evidence of a specializatiftn and corresponding 

 limitation of the feeding habit— the increased efficiency of one structure 

 (•orrcspoiKling to llie increased importance to the fish of the related kind 

 of food, and tlie (iefective development of the correlated structure indicat- 

 ing an abandoumenl of the food for whose apjjropriation it was especially 

 fitted. On the other hand, the absence of these inverse correlations marks 

 an omnivorous hal)it— as in the catfishes. whose jaws, teeth, gill-rakers and 

 pharyngeals are all moderately developed, while the food is correspondingly 

 indiscriminate. 



