1902.] accompanying Mechanical Disturbance in Metals. 287 



It will also be noticed that whereas recovery is complete in 1 minute 

 when the vibration amplitude is small, it is not quite complete within 

 that time when the vibration amplitude is large. Greater strain pro- 

 longs the period of recovery. Owing to want of complete recovery, 

 the base line is tilted slightly upwards. This slight displacement does 

 not materially affect the results, provided the shifting is slight. From 

 other records taken through a greater range of stimulation, it appears 

 that in a curve obtained with electromotive variations as ordinates 

 and amplitudes of vibrations as abscissa?, the first part of the curve is, 

 generally speaking, slightly convex to the abscissa (the convexity is 

 pronounced when feeble stimulation gives negative response), it is then 

 straight in the middle and concave in the last part. A limiting deflec- 

 tion is approached with high amplitude of vibration. The shape of 

 the curve is modified by the molecular condition of the wire. 



Maximum Effect. 



If instead of a single vibration of a given amplitude we superpose a 

 rapidly succeeding series, the individual effects are added up and a 

 maximum deflection is produced which remains practically constant as 

 long as the vibration is maintained. (A single ineffective stimulus 

 may thus become effective by the additive effect of several.) Too 

 long-continued vibration may cause fatigue, but during half a minute 

 or so, the maximum effect is very definite (in tin). For a definite 

 amplitude of vibration there is a definite maximum, which increases 

 with the amplitude. For example, a single vibration of 5° gave a 

 deflection of 3*5 divisions, the same when continued at the rate of 

 four times per second gave a maximum deflection of eighteen divisions. 



Again, with the same wire, a single vibration of 10° gave a deflec- 

 tion of 4*5 divisions, but continued vibration gave the definite maxi- 

 mum of 37*5 divisions. I give below a curve (fig. 10) which shows 

 the maximum effect for different amplitudes of vibration. 



Hysteresis. — Allusion has already been made as to the increased 

 sensitiveness conferred by preliminary vibration. Being desirous of 

 finding out in what manner this is brought about, I took a series of 

 observations for an entire cycle, that is to say, a series of observations 

 were taken for maximum effects, starting from 10° and ending in 100°, 

 and backward from 100° to 10°. Effect of hysteresis is very clearly 

 seen (fig. 10, A) ; there is a considerable divergence between the forward 

 and return curves, the return curve being higher. On repeating the 

 cycle several times, the divergence is found to be very much reduced, 

 the wire on the whole is found to assume a more constant sensitiveness. 

 In this steady condition, generally speaking, the sensitiveness for smaller 

 amplitude of vibration is found to be greater than at the very beginning, 

 but the reverse is the case for stronger intensity of vibration. 



