612 DR W. PEDDIE ON 



Modification of the Apparatus. 



In the former experiments, an iron wire 89"1 cm. in length, and 0*101 1 cm. in 

 diameter, was used. Before measurement, the ends of the wire had been soldered 

 into holes drilled axially in brass rods. One of these rods, with the wire suspended 

 from it, was clamped in a vertical position. To the other rod, a lead ring, of consider- 

 able moment of inertia, was attached symmetrically. The oscillations were produced 

 by twisting the wire in opposite directions alternately, by hand, the impulses being- 

 timed to suit the natural period of the wire, until the required maximum was reached. 

 The unavoidable pendulum-swings of the wire were then damped out by hand as 

 rapidly as possible. 



The same wire was used in the new series of experiments, but the upper rod was 

 passed through a cylindrical hole, of the same diameter as the rod, drilled vertically 

 through a fixed metal block. The rod ended in a head which rested on the upper 

 horizontal face of the metal block and thus supported the oscillating system. A 

 horizontal lever, attached to the head, resting normally in contact with a stop, could 

 be turned out from, and back to, the stop through a considerable angle. In this way, 

 large or small oscillations could be readily produced. Two such double motions of 

 the lever, properly timed, were found to be sufficient to produce the largest oscillation 

 desired. 



The angle of oscillation was read off on a scale, attached to the circumference of 

 the lead ring, by means of a telescope placed a short distance away. Formerly, the 

 reading was got by means of a fixed pointer placed close in front of the scale ; in the 

 new observations, it was got by means of a fibre placed in the focus of the telescope. 



In the former experiments, the lead ring was of such moment of inertia that ten 

 complete oscillations were performed in about 80 seconds. In the new series, for a 

 reason which will appear subsequently, a lead ring of the same weight but of greater 

 moment of inertia was used. When tested during the course of the first experiment 

 (12.7.94), and subsequently at the date 25.7.94 after a number of experiments had 

 been performed, nineteen complete oscillations took place in 300 seconds. 



Symmetry of the Oscillations. 



In the former experiments, positive elongations alone were read ; for the period 

 of oscillation was too short to admit of both elongations being read with accuracy. 

 The zero from which the elongations were reckoned was the point about which the 

 oscillations took place when the motion had died down to a large extent. This is a 

 point about which the oscillations must be symmetrica], should such a point of 

 symmetry exist throughout the whole motion. But the existence of such a point 

 cannot, from the nature of the problem, be postulated a priori. The first positive 

 elongation causes a considerable positive set of the zero from its original j>osition. 



