RESPIRATION OF FISHES. 117 



The gymnoti, like our eels, are fond of swallowing and 

 breathing air on the surface of the water ; but we must not 

 thence conclude that the fish would perish if it could not 

 come up to breathe the air. The European eel will creep 

 during the night upon the grass ; but I have seen a very 

 vigorous gymnotus that had sprung out of the water, die on 

 the ground. M. Proven9al and myself have proved by our 

 researches on the respiration of fishes, that their humid 

 bronchia) perform the double function of decomposing the 

 atmospheric air, and of appropriating the oxygen contained 

 in water. They do not suspend their respiration in the 

 air; but they absorb the oxygen like a reptile furnished 

 with lungs. It is known that carp may be fattened by being 

 fed, out of the water, if their gills are wet from time to 

 time with humid moss, to prevent them from becoming dry. 

 Fish separate their gill-covers wider in oxygen gas than in 

 water. Their temperature however, does not rise ; and they 

 dve the same length of time in pure vital air, and in a 

 mixture of ninety parts nitrogen and ten oxygen. We 

 found that tench placed under inverted jars filled with air, 

 absorb half a cubic centimetre of oxygen in an hour. This 

 action takes place in the gills only ; for fishes on which a 

 collar of cork has been fastened, and leaving their head out 

 of the jar filled with air, do not act upon the oxygen by the 

 rest of their body. 



The swimming-bladder of the gymnotus is two feet five 

 inches long in a fish of three feet ten inches.f It is sepa- 

 rated by a mass of fat from the external skin ; and rests upon 

 the electric organs, which occupy more than two-thirds of 



specifically different. The Indians mentioned to us a black and very 

 powerful species, inhabiting the marshes of the Apure, which never 

 attains a length of more than two feet, but which we were not able to 

 procure. The raton of the Rio de la Magdalena, which I have described 

 under the name of Gymnotus aequilabiatus (Observations de Zoologie, 

 vol. i.) forms a particular sub-genus. This is a Carapa, not scaly, and 

 without an electric organ. This organ is also entirely wanting in the 

 Brazilian Carapo, and iu all the rays which were carefully examined bf 

 Cuvier. 



f Cuvier has shown that in the Gymnotus electricus there exists, 

 oesides the large swimming-bladder, another situated before it, and much 

 smaller. It looks like the bifuicated swimming-bladder in the Gymnotui 



