Units in the Electromagnetic Unit. 435 



magnet by varying the distance of the magnet, and found the 

 results to agree very closely with one another. The mean 

 value of H obtained with one magnet is '15955; and the mean 

 value obtained with the other is *15937; so that the mean of 

 these two is 



H= -15947. 



The formula used in the calculation of H is 



2tt AW 



H= 



v, 



t(k-l)(k + l)V tan^ ; 



where t is the period of vibration of magnet under H, k the 

 distance of the centre of the magnet from the magnetometer, 

 I half the length of magnet, i the moment of inertia of the 

 magnet, and </> the angle of deflection of the magnetometer. 



We have now come to the evaluation of " v." The formula 

 (2) gives 



e = l-01172 xlO 8 (C.G.S.) electromagnetic units. 



Hence 



v = 294-4 x 10 8 centims. per second, 



which agrees well with the latest value obfeined by Sir Wil- 

 liam Thomson, namely 293 x 10 8 . 



Although I took as much care as possible in making all the 

 above measurements leading to this evaluation of " v" yet 

 since, from want of time, it was only on one occasion that I 

 was able to make the complete measurements, there may have 

 been some cause or causes of error unnoticed. I intend there- 

 fore to repeat the whole experiment, and hope to be able to 

 make a further communication. 



In conclusion, I must say (and I say with extreme grati- 

 tude) that if the experiment be in any way satisfactory, it is 

 chiefly due to the very able and kind instructions given me 

 by Sir William Thomson and his assistants in carrying it 

 out. 



Addition, Nov. 18, 1880. — These experiments have, since 

 the communication of the above paper to the British Associa- 

 tion, been several times very carefully repeated, with in every 

 case a confirmation of the close accuracy of the determination 

 of the electrostatic value of electromotive force. In the elec- 

 tromagnetic determination, however, a correction has been 

 made for the torsion of the single silk fibre by which the 

 needle of the tangent-galvanometer was suspended. As it 

 was supposed that the torsion of a single fibre of silk might be 

 neglected, no correction was made in the first results ; but the 



