1, MEMOIRS OF THE CARNEGIE MUSEUM. 
The family Gymnotide, as discussed in this monograph, includes all of the 
species of the two families, the Gymnotide and the Electrophoride as restricted by 
Gill. The electric eel, Electrophorus electricus Linnzus, has been included in this 
family for two reasons. Its affinities with Gymnotus carapo Linnzeus are very 
close, and it is more closely related to the other Gymnotids than to any other 
group of fishes. The relation of H#. electricus to G. carapo is shown by the following 
comparison. 
I. CxHaracters Common To BotH GENERA. 
Depressed head; body subeylindrical and elongate; teeth large, in one or two 
rows in each jaw, conical, in sockets; lower jaw slightly projecting; eyes small; 
no frontal fontanel; parietal fontanel small and almost covered by the overhanging 
occipitals; posterior air-bladder long and conic; origin of anal fin just below tips 
of pectorals; anus below gill-opening. 
II. Cuaractrers RESTRICTED TO ELECTROPHORUS. 
Anal turned up so as to form a false caudal; scales wanting; electric organs 
well developed. 
III. CwHaracters RESTRICTED TO GYMNOTUS. 
A small caudal appendage projecting beyond the anal fin; no electric organs, 
or at least only indications of Hunter’s organs; scales present. 
It will be seen that the presence of electric organs is the point of largest differ- 
ence between Electrophorus and Gymnotus and as pseudo-electric organs are known 
for other species of the Gymnotide, it does not seem that EHlectrophorus should 
stand in a separate family. Plate XVI shows two views of the skull of G. carapo. 
That of Electrophorus is the same in almost every detail, except that it is more 
depressed. 
Three subfamilies are recognized. The Gymnotine just discussed, the Sterno- 
pygine, and the Sternarchine. The last two named differ from the first especially 
in two particulars: they are compressed and have both frontal and parietal fon- 
tanels. The Sternopygine differ from the Sternarchine in the absence of a caudal 
fin. These two subfamilies parallel each other quite closely in their variations. 
Plate XV shows an outline of the head of a typical species of each genus of the 
family. The parallelism of the Sternopygine and Sternarchine is particularly 
evident in the development of long-snouted forms, short-snouted forms, and 
toothless forms. Plates XVII and XVIII show the modification of the skull in the 
long-snouted Rhamphichthys rostratus as compared with the short-snouted Higen- 
mannia virescens, both fishes being of the subfamily Sternopyginw. These plates 
may be compared with Plate XV as regards the presence or absence of the frontal 
fontanel. 
soulenger (Archiv fiir Naturgeschichte, Jahrgang 1904, Bd. I, Heft 2) con- 
siders the Gymnotide as an offshoot from the Characide.. The Gymnotide seem to 
