440 Mr. Norman Shaw on Increased Accuracy 



actual process of the problem. A comparison of (13) and 

 (15) shows that the former equation is much more cumber- 

 some to deal with than the latter, and as equation (8) is so 

 compact, and holds over that range for which it is necessary 

 in practice for the correction of deflexion readings, we consider 

 it therefore preferable. 



Methods of Observation. 



The suspension with which these experiments have to deal 

 will carry a current of half an ampere without changing the 

 period of vibration of the supported system by more than 

 one part in 50,000, and will remain constant for weeks to 

 the same degree of accuracy. The following observations 

 on the period, which were recorded on a chronograph 

 controlled from the McGrill Observatory, illustrate this 

 constancy : — 



Period. 



T. No current flowing 11*6672 seconds 



Current of half an ampere . . . 11*6669 „ 



No current flowing 11*6669 ,, 



Current of half an ampere . . . 11*6669 ,, 



II. (Slight change in weight of system) 



No current flowing 11*6571 „ 



Current of half an ampere... 11*6574 „ 



No current flowing 11*6)75 „ 



Current of half an ampere . . . 11*6573 „ 



III. May 5, 1910 11*6566 „ 



„ 28, „ 11*6564 „ 



June3, „ 11*6564 „ 



„ 13, „ 11*6564 „ 



IV. (Moment of inertia changed, but no change 



in weight of system.) 



Aug. 2, 1910 6*5659 „ 



Nov. 14, 1910 6*5652 „ 



The elastic fatigue produced by the continuous torsional 

 oscillation* in the determination of these periods was not of 

 a sufficient magnitude to appreciably affect these observations 

 in any way. 



The deflexions were produced by passing a current through 

 the dynamometer. This current passed through a constant 

 resistance in an oil-bath, and it was arranged that the drop 



* 



Kelvin, Eneycl. Brit. Art. " Elasticity," § 30. 



