2(3 Messrs. Garrett and Lucas on a 



and were much brighter than those obtained from the electric 

 spark; but assuming as before one fifth of a band's breadth 

 as the minimum displacement which would be visible to the 

 eye, we have, since d = 4.5 miliim., 



u = 200 metres per second 



as minimum velocity of the eether in the cell which would be 

 detected. The strength of the magnetic field was 300 (J.G.IS. 

 units. 



In our final experiment, while retaining the electromagnet 

 in position, we removed the electrolytic cell and set up in its 

 place a horizontal electrostatic field by means of a condenser, 

 consisting of two tin plates \\ centim. apart. These plates 

 measured 17 centim. x 12 centim., but one of them was cut 

 at the middle of its length so as to pass between the pole-pieces 

 of the electromagnet. The position of the apparatus was there- 

 fore this : — The plates of the condenser were in vertical planes 

 parallel to the direction of the beams of light. Between these 

 plates were the poles of the electromagnet, the one vertically 

 above the other. The plates were connected with the poles 

 of a Wimshurst machine, and also with the terminals of a 

 Kelvin electrostatic voltmeter. The negative pole of the 

 machine was connected further with the case of the voltmeter 

 and led to earth. The positive plate of the condenser was 

 insulated on glass plates. 



The strength of the magnetic field was the same as before, 

 300 C.G.S. units ; and between the plates of the condenser 

 the difference of potential was varied, the maximum being 

 6(J00 volts. During each set of observations the potential 

 was kept constant ; but in none of the observations was any 

 motion of the interference-bands discerned. 



Cavendish Laboratory, 

 Cambridge. 1897. 



IY. A Nickel Stress Telephone. By T. A. Garrett, M.A., 

 and William Lucas, M.A. * 



IT is well known that, if a magnetized nickel wire be 

 subjected to longitudinal stress, variations of this stress 

 produce variations in the magnetization of the wire. If the 

 stress be a pressure, an increase in the pressure produces an 

 increase in the magnetization, and a decrease in the pressure 

 a decrease in the magnetization. 



Suppose then that a magnetized piece of nickel wire, 



* Communicated by tbe Physical Society: read April 9, 1897. 



