782 



THE AMERICAN NATURALIST. [Vol. XXXVII. 



isolated lakes which have never been visited. It is known that 

 some of these lakes, as Patzcuaro and Zirahuen, have in them 

 one or more characteristic species and no doubt most of them 

 have. The Lerma river system is far from being thoroughly 



characteristic as if it were an island in the sea. 



All of the many beautiful lakes in this area, now isolated, 

 evidently at one time drained into the Lerma, and so became 

 stocked with fishes. It is often argued that fishes become estab- 

 lished in isolated lakes by their eggs being carried accidentally 

 by water birds. Although while these birds are feeding some 

 eggs of fishes might cling t<> their feathers or legs and be taken 

 to adjacent waters, yet I much doubt the dispersion of fishes in 

 this way. Were this method of dispersion at all common fishes 

 would surely have been found in Shoshone and Lewis lakes in 

 the Yellowstone Park ; moreover in the isolated lakes in the 

 valley of the Lerma the viviparous fishes are about as evenly 



The Rio Balsas is one of the largest rivers in Mexico. It is 



and two of killifishes, are known from the Balsas, and yet these 

 two families comprise nearly two thirds of the fishes of the Lerma 

 basin. The Balsas is far from being thoroughly explored; 



fauna and to indicate that it contains comparatively few species 

 of fishes. 



In Mexico there are four quite distinct fish faunas, and though 

 they overlap at the borders, the map may fairly indicate where 

 each fauna prevails. 1 The origin of these fish may be approxi- 

 mately given as follows: from the Colorado river 9, from the 

 Rio Grande 80, from the Lerma 49, and from Central America 

 about 108; total 246. 



The fish fauna of northern Mexico is essentially that of the 

 Rocky Mountains and eastern United States, or that part of 

 the United States adjacent to Mexico. This eastern fauna has 



