THE GYMNOTID EELS OF TROPICAL AMERICA. : 175 
inferior. The snout of S. miilleri is also straight and is only slightly out of line 
with the long axis of the body. It is, however, directed downward to some extent 
and the mouth is more or less inferior. All three species are thus adapted to feed 
upon the bottom fauna with the minimum of effort. It seems probable that they 
feed by sucking up quantities of mud and food after a suitable feeding-place has 
been found. Forms like these taken from the stomachs are usually found in large 
numbers, close together. 
No stomachs of the long-snouted Gymnorhamphichthys hypostomus Ellis were 
examined. 
General Considerations. 
Two factors control the nature of the food taken by any fish, namely, (1) the 
structure, and (2) the size of the fish. In the Gymnotide the only structure that 
needs consideration is the mouth. The other structures which might influence the 
selection of food, such as shape, nature of fins, and the like, are all held in common 
by the several species of this small family. Considering size and mouth there are 
four types of Gymnotide, (1) large, large-mouthed specimens (adult only); (2) small, 
large-mouthed (young of large-mouthed); (3) large, small-mouthed eels; and 
(4) small, small-mouthed eels. The last three are ecologically the same, since 
they all have small mouths. Class two, however, differs in that the members of 
this group ultimately by growth attain to the first class. G. carapo, EF. electricus, 
S. macrurus, S. albifrons and S. brasiliensis are the large-mouthed species of this 
family. 
They are the only species examined which had eaten either fish or freshwater 
shrimps, or both. These two items, which are the most bulky food taken from 
Gymnotid stomachs, were found only in the largest specimens of the species be- 
longing to the first class. 
A comparison of the various tables shows that the young of all species partake 
of much the same food. They feed upon entomostraca, the larvee of insects and 
small annelids. The second, third, and fourth classes are therefore alike as regards 
the food taken. There remains, however, the ability of the second class to change. 
Forbes has found that the food of all small North American fishes is much the 
same, being for the most part entomostraca and the larve of insects. For the 
predaceous fishes those which as adult feed largely upon other fishes he has also 
shown a regular cycle of foods from the young to the adult. Beginning with 
entomostraca and insect larvee they pass to Annelida and adult insects, small 
Crustacea, large Crustacea, and finally to other fishes. Precisely this progression 
is shown in the first two tables (pp. 168 and 170). On the other hand, the non- 
predaceous members of this family, which are limited by small mouths, pass only 
from entomostraca to insect larve. 
Schlesinger divides the species of this family into three groups. The first 
eroup contains H. electricus and G. carapo. He imagines that the food of these 
two species must be the same because of their general resemblance. This was 
