Brown and Smith — Subsidence of Torsional Oscillations. 219 



value as the direct magnetic field, for a period of three minutes ; the alternating 

 field was then switched off, and the vibrator and wire brought to rest. Direct 

 currents of the same value as formerly were now put through the solenoid 

 and wire, and the steady deflection obtained on the scale was only one 

 niilUnietre, which showed that some decided change had taken place in the 

 wire, which, for want of a better name, we have called temporary magnetic 

 fatigue. That tliis effect is temporary is shown by the fact that it can be 

 cured in several ways — (1) by putting an alternating magnetic field round the 

 wire of lower frequency than that which caused the fatigue ; then taking off 

 the alternating field, and allowing the wire to rest for some time, it slowly 

 recovers ; in one case, however, we observed that it took seventeen hours to 

 fully recover : (2) another way we found effective was to relieve the wire of 

 its load, and send a direct current round the solenoid ; this made the wire 

 recover at once. The above was when the load on the wire was 10° grammes 

 per sq. cm. 



When the small load of 0'5 x 10° grammes per sq. cm. was on the wire, 

 and an alternating magnetic field of 13 c.g.s. units and frequency of 140 per 

 second was round the wire for about three minutes, and then tested, the wire 

 was found to be about half fatigued ; and when the larger load of 1'5 x 10' 

 grammes per sq. cm. was on, and au alternating field of 20 c.g.s. units of 

 frequency 140 per second for three minutes, again, on testing the wire, it was 

 found to be about three-quarters fatigued.' This magnetic fatigue does not 

 occur in soft nickel wires as shown below. We have, however, made 

 arrangements for investigating further and more fully this property of 

 hard nickel wires. 



The wire was now taken down and made as soft as possible by the 

 following means : — it was suspended in a vertical position and allowed to 

 hang loosely under its own weight only, and raised to a bright cherry-red 

 heat by means of a broad Bunsen burner by heating it from the top down- 

 wards three times in succession, and when cooled its simple rigidity was 

 about 708 x 10*^ grammes per sq. cm. The wire was again put into position 

 inside the solenoid, and the small load 0-5 x 10° grammes per sq. cm. hung on 

 the lower end ; the damping of the torsional oscillations was observed as 

 before, (1) when the wire was under the influence of the Earth's vertical force, 



(2) when a longitudinal magnetic field of 13 c.g.s. units was round it, and 



(3) when alternating magnetic fields of V6 c.g.s. units each and of fre- 

 quencies 20, 30, 50, 100 per second were round the wire. The results 

 obtained are shown in Table II. 



1 This means that if the steady deflection is 40 divisions on the scale when the wire is fresh, the 

 deflection when half fatigued is (1 - ^) 40 = 20 divisions ; similarly when the wire is three-quarters 

 fatigued the deflection on the scale is (1 - f) 40 = 10 divisions ; in the same way we get ( 1 - 1) 40 = 

 for total fatigue. 



