1294 
of frequency. Indeed, we can infer from the above value of g, that 
the error made on this assumption is about 11°, i.e. less than +; 7. 
The method, by which the phase was determined, was however 
not so accurate, that a difference of this amount could be observed. 
While L followed the oscillations of the cross-wire over the scale, 
a second person looked at the pendulum and gave a sign every 
time when it passed through the position of equilibrium in such a 
way that a positive reversal of the current took place. Now I ob- 
served, where the wire stood at these moments and in what direc- 
tion it was moving. If we had to do with effect 1, the wire should 
have its greatest positive deviation at the moment of a sign. For 
we have seen, that at the instant of a positive reversal, the first 
effect produces an impulse in negative direction. We then found for 
the term in the Fourtmr-series that is of importance here 
Qn 
— A COS 7 t 
(a = — being positive). From this and (7) we infer, that, if p is 
7 2 ) 
neglected, w has its greatest positive value for ¢= 0. 
This result is confirmed by a simple reasoning. If external forces 
shall cause a body to oscillate more rapidly than it would under 
the action of the elastic force alone, they must drive it towards 
the position of equilibrium. Now at the instant of a positive reversal 
the effect we want to observe consists in an impulse in negative 
direction. If therefore the cylinder shall be driven towards the position 
of equilibrium, it must be on the positive side of it. 
As has been said already effect 3 (a, b) has the same or the 
opposite phase as 1; if in the latter case it exceeded the first effect, 
we should see the cross-wire just at the negative end of its oscil- 
lation at the moment of a sign. 
As to effect 2, this differs a quarter of a period in phase from 1; 
if it exists alone or preponderates, as was always the case before 
the application of compensation II, the wire will be in its position 
of equilhbrium at the moment we consider, the direction of its 
motion depending on circumstances. 
If all the effects exist at the same time an intermediate phase 
will be observed. 
§ 8. Description of the experiments of September 4%. Intensity 
of the current in the coil S 29,5 milliamperes. Every time before 
the determination of the deviation and the phase I waited during 
80 to 100 full periods in order that all free oscillations might be 
extinct and the cylinder might perform its forced vibrations only. 
