ANOMALOUS BEHAVIOUR OF DELICATE BALANCES, ETC. 



393 



conducted. The results obtained when the loads were respectively 0, 10, and 

 50 gr. are shown graphically in Fig. II. They are typical of all the other loads 

 employed. 



Loa d = 0. 



o a. d=l grm s. 



at 



a: 



"-.._ L. o a.d=50 yrrns. 



Vi b rations. 



The sinuous and dotted line curves represent the fluctuations in the restiiig-point 

 (R.P.) ; the others indicate very approximately the mean rate of change in the R.P., 

 from a maximum to a somewhat lower and normal value. In these particular cases 

 we observe that the variations in the R.P. are not so strongly marked as are those 

 which we have represented in Fig. I. 



The dotted line graphs suggest, more or less markedly, curves of sines. We also 

 note that an increase in the load produces in the first portions of the curve a greater 

 degree of steepness. But in these, as in some hundreds of similar cases which were 

 observed, the R.P. ultimately tends to assume a constant and ideal value. 



When the decay in the amplitude of the oscillations has become so great that 

 reliable readings of the pointer are no longer possible, one of the pans may be gently 

 tapped with its arrester, and an additional set of observations taken. Plotting the 

 curve as before, we find that it still continues sinuous in form, but that the axis of 

 the sinuosities now travels almost horizontally. If, however, the experiment is 

 repeated after the beam has been arrested for a short period then the results obtained 

 are similar to those shown in Fig. II. 



VOL. cox. A. 3 E 



