The Gymnotus experimented with : — Collectors employed. 361 



ciety, and I give them as necessary preliminary experiments 

 to the investigations which we may hope to institute when the 

 expected supply of animals arrives (1752.). 



1755. The fish is forty inches long. It was caught about 

 March 1838; was brought to the Gallery on the 15th of 

 August, but did not feed from the time of its capture up to 

 the 19th of October. From the 24th of August Mr. Bradley 

 nightly put some blood into the water, which was changed for 

 fresh water next morning, and in this way the animal per- 

 haps obtained some nourishment. On the 19th of October it 

 killed and eat four small fish ; since then the blood has been 

 discontinued, and the animal has been improving ever since, 

 consuming upon an average one fish daily*. 



1756. I first experimented with it on the 3rd of September, 

 when it was apparently languid, but gave strong shocks when 

 the hands were favourably disposed on the body (1760. 1773, 

 &c.). The experiments were made on four different days, 

 allowing periods of rest from a month to a week between 

 each. His health seemed to improve continually, and it 

 was during this period, between the third and fourth days of 

 experiment, that he began to eat. 



1757. Beside the hands two kinds of collectors were used. 

 The one sort consisted each of a copper rod fifteen inches 

 long, having a copper disc one inch and a half in diameter 

 brazed to one extremity, and a copper cylinder to serve as a 

 handle, with large contact to the hand, fixed to the other, the 

 rod from the disc upwards being well covered with a thick 

 caoutchouc tube to insulate that part from the water. By 

 these the states of particular parts of the fish whilst in the 

 water could be ascertained. 



1758. The other kind of colleclors were intended to meet 

 the difficulty presented by the complete immersion of the fish 

 in water; for even when obtaining the spark itself I did not 

 think myseH" justified in asking for the removal of the animal 

 into air. A plate of copper eight inches long by two inches 

 and a half wide, was bent into a saddle shape, that it might 

 pass over the fish, and inclose a certain extent of the back 

 and sides, and a thick copper wire was brazed to it, to convey 

 the electric force to the experimental apparatus; a jacket of 

 sheet caoutchouc was put over the saddle, the edges project- 

 ing at the bottom and the ends ; the ends were made to con- 

 verge so as to fit in some degree the body of the fish, and the 

 bottom edges were made to spring against any horizontal sur- 

 face on which the saddles were placed. The part of the wire 

 liable to be in the water was covered with caoutchouc. 



* The fish eaten were gudgeons, carp, and perch. 



