, 382 Electro-Magnetic Experiments. 
ence an experiment I cut the same circuit of wire again near one of 
its ends and soldered into it a disk of bismuth about three fourths of 
an inch in diameter, this forming a compound circwit. 1 then appli- 
ed the lamp near the twisted joint as in No. 2. The effect was the 
same as before the bismuth had been inserted. 
5th. Applying my finger to the solder between the bismuth and 
the copper of the compound circuit, the needle was put into such mo- 
tion by the electricity generated by the animal heat that in three 
seconds, it acquired a velocity sufficient to turn it several times round. 
It was impossible to touch this soldered joint for the fourth of a sec- 
ond without putting the needle in motion. 
6th. Finding this compound circuit to be so sensible a differential 
thermometer, I took the opportunity to determine the progress of 
heat through the wire. The following table shows the result of the 
application of the lamp at different distances from the bismuth. 
The adjacent table shows that the times in which the heat moves 
along the wire are nearly as the squares of the distances. This ap- 
pears from 2 and 4, and 3 and 9, in the first and second column; 
and from 4 and 16, 5 and 24 (25), 6 and 36, 7 and 50 (49)in the 
first and third columns. 
Time from the application of the heat 
Distance till the needle began to move. till the needle had moy- 
Inches. Min. Sec. ed 6°. 
18 4 0 No motion. 
12 3 0 slight agitation. 
9 ) 45 75 seconds. 
7 0 39 OO is 
6 0 25 DO! Cone 
5 0 16 24 6 
4 0 12 BG: 088 
3 0 9 Lea 
2 0 4 Or 
My only object, however, in these experiments was to determine 
what length of uniform uninterrupted metal must intervene, between 
the point where 1 applied the heat and the connexion with the gal- 
vanometer and also the length of time which I might occupy in the 
operation. My conclusion was that the metal under experiment must 
extend in form of a wire, one sixteenth of an inch in diameter, at 
least twelve inches from the part intended to be heated, and the 
heating and experiment must then occupy less than three minutes. 
Having procured several metals in suitable forms I proceeded to 
my main experiments. Heating one piece of the metal in the first 
