Transparent Bodies by the action of Magnetism. 493 



the angle through which the whole system is turned be suffi- 

 ciently small. By increasing the distance from the scale to the 

 mirror, the precision of measurement is in some degree indefi- 

 nitely increased, provided the telescope have a magnifying power 

 sufficient to show the divisions clearly. In my experiments the 

 most convenient distance appeared to me to be that of l m, 25 ; 

 an angular displacement of 80 seconds corresponded to each 

 division of the graduated scale ; and since a quarter of a divi- 

 sion could be easily estimated, the accuracy of the measurement 

 reached twenty seconds. 



In my instrument, the magnetized needle was a large bar of 

 steel, m, 3 long by m '015 in diameter; the wire of the galvano- 

 meter was 100 metres long and 1 millim. in diameter, and was 

 rolled on an oval frame of yellow copper, m, 15 in length and 

 m, 10 in width. The metallic mirror was a square of m, 04 each 

 side, and the cocoon thread without torsion was m- 35 long. In 

 the interior of the frame of brass was another frame of red copper, 

 of the same form but 1 centimetre thick, intended to damp the 

 oscillations of the needle by the counter-action of the induced 

 currents which the movement of the needle developes in its mass. 

 In the construction of this apparatus I employed M.Ruhmkorff, 

 who acquitted himself with his usual ability. 



If along the wire of such a galvanometer a current of very 

 short duration is passed, this cuiTent communicates to the needle 

 an impulse proportional to the integral of the actions which it 

 exercises during the successive instants of its duration, and con- 

 sequently proportional to the total quantity of electricity which 

 it causes to pass through any section whatever of the wire ; this 

 quantity is, as was seen above, the measure of the magnetic action 

 exerted at the point where the helix of induction is situated. 

 If no resistance was offered to the movement of the needle, it 

 would perform oscillations entirely comparable to those of a pen- 

 dulum in vacuo, and the sine of the semi-deviation would exactly 

 measure the initial impulse. In reality the needle suffers various 

 resistances, among which the most important is that which pro- 

 ceeds from the reaction of the induced currents developed in the 

 copper frame and in the galvanometric wire itself; and the am- 

 plitude of its oscillations decreases so rapidly, that the previous 

 relation does not take place. But if the decrease in the oscilla- 

 tions occurs in geometrical progression, and if the initial devia- 

 tion do not exceed a certain limit, it may be demonstrated, ac- 

 cording to M. Weber, that the initial displacement of the image 

 of the scale is proportional to the initial impulse. 



I assured myself by repeated trials that these conditions were 

 satisfied in my galvanometer. I took, in consequence, us a mea- 

 sure of the induced current and of the magnetic action, the dis- 



