4 Mr. W. Sturgeon on the Thermo- Magnet ism of 



point ; it is also inclosed in a box with a glass cover. Care 

 was taken to neutralize the metallic frame between every two 

 experiments, by plunging it into cold water. 



13. Fig. 2. will represent this frame of bismuth. It was 

 heated successively at the points a, b, c, d, close to the angles ; 

 but those angles were kept out of the flame of the spirit-lamp. 

 These points were selected for the points of heat, from a no- 

 tion which I had previously entertained, that heat might pos- 

 sibly be obstructed by turning sharp angles, and thereby in- 

 fluence the direction of the electric currents to which it gave 

 birth. 



fa, the current flowed from 6 to a 



When the point I 6, * b to a 



of heat was at I c, i to a 



\^d, a to b. 



To prevent confusion, the side a b, only, is chosen to show 

 the directions of the electric currents. It is to be understood, 

 however, that those currents were continuous round the rect- 

 angular frame. 



1 4. When the experiments were repeated, the direction of 

 the current changed when the point of heat was at b; at all 

 the other points the currents proceeded as at first. By vary- 

 ing the situation of the point of heat several times near to b, 

 it was found that, when applied anywhere within half an inch 

 of the angle, the current was from 6 to a as at first ; but when 

 the point of heat was one inch distant from the angle b, the 

 current invariably proceeded from a to b. From these re- 

 sults it was evident, that between half an inch and one inch 

 from the angle, there was a point which, if heated, no electric 

 current would be excited. By various trials this neutral point 

 was found to be situated at a little more than half an inch 

 from the angle b. So that in general, if all the first half-inch 

 were heated, the current would proceed from b to a; but if 

 the more distant half-inch from the angle was heated, the 

 current flowed in the opposite direction. Again, as this last 

 current was also in the opposite direction to that excited by 

 heating the point a, there would evidently be another neutral 

 point still nearer to a. This point was determined at nearly 

 half way between the angles a and 6, but a little nearer to 

 the former than to the latter. 



15. In this way the situation of the point of heat was varied 

 in the side c d. One neutral point only on this side was de- 

 tected, which was nearly half-way between the angles c and d. 

 If any point in, or the whole of the half nearest to, c were 

 heated, the current proceeded from dto c; but if heat was 

 applied to the other half nearest to d, or to any particular 

 point in that half, the current flowed from c to d. 



16. By 



