STRAIN ON THE PHYSICAL PROPERTIES OP MATTER. 
821 
terminated at its lower extremity S in a ring, through which a pin pp could just 
pass. The pin pp served to connect the brass box A with the bar of the vibrator by 
passing through two rings r, r, attached to the top of A, and placed one on each side 
of the ring at S. In this way the box A was well secured to VY, and without the 
least chance of any shifting taking place, as the wire vibrated torsionally. Two 
smaller boxes B, B, were suspended from W, and could be clamped, as shown in the 
figure, to any part of it. All three boxes were provided with hinged doors * of the 
same height as the boxes, so as to admit of the introduction of two sets of cylinders 
C and L, the one set of brass and the other of lead ; these doors, after the cylinders 
had been arranged in the boxes, were kept shut by means of springs and clasps. The 
mass of each of the brass cylinders C was 100 grms., and that of each of the lead 
cylinders L 2000 grms. The masses of each of the cylinders in the two sets were 
adjusted by myself, so as to agree with some standard masses within 1 part in 
10,000. Each of the cylinders L had been cast in the same mould, and the 
dimensions had been carefully measured by means of calipers reading to xoVoth inch. 
The moment of inertia of each cylinder was determined from the dimensions and 
mass, and afterwards by certain indirect methods, which last, when suitable pre¬ 
cautions were taken, gave results in good accordance with those which had been 
obtained by the direct method. Moreover, the moments of inertia of all the lead 
cylinders were found to be very nearly equal to each other. The mode of proceeding 
was as follows :—Eight of the lead cylinders L were placed in the box A, and on the 
top of these was put a wooden cylinder W, fitting rather tightly in A. On the top 
of W, and fitting into a circular cavity let into the centre of it, were placed sixteen 
of the brass cylinders C, arranged in sets of four each, each set having its four 
constituents bound together by an elastic band d. The w'ooden cylinder W served 
not only to prevent any rocking of the lead cylinders as the wire vibrated, but also to 
enable the experimenter to readily adjust the brass cylinders C in their proper 
positions in the box A. 
When all the necessary adjustments had been carefully completed, the whole was 
allowed to remain for five days at rest, except that the wire was caused to vibrate 
several hundred times on each day ; finally, the logarithmic decrement and the 
vibration-period were determined from a very large number of oscillations. As soon 
as these had been determined in a sufficiently satisfactory manner, two of the lead 
cylinders L were removed from A, and four of the little brass cylinders C were trans¬ 
ferred from A to the boxes B, B, two being placed in each box. The boxes B, B, had 
been at the outset so placed on V V that the above transference would exactly make 
up for the loss of the moment of inertia which would follow from the removal of the 
two lead cylinders. After this second adjustment and a rest of twenty-four hours, 
followed by a large number of preliminary vibrations, the logarithmic decrement and 
the vibration-period were again determined. Again, two more lead cylinders were 
* Not shown in the figure. 
