48 THE ROYAL SOCIETY OF CANADA 



by thin asbestos paper and are cemented in place with a mixture of 

 sodium siHcate and powdered talc. The frame work carrying the coils 

 and supporting the tube is mounted on a brass plate provided with 

 levelling screws. Thus the cylinder may be centred in the tube 

 either by these screws or by means of the jacks which support the 

 entire apparatus. 



Various methods of observing the oscillation of the cylinder were 

 tried. A mirror on the cylinder was found to be impracticable owing 

 to the distortion due to irregularities in the glass tube. In fact any 

 horizontal motion is difficult to measure accurately. Finally the 

 following method was adopted. After the gold cylinder was com- 

 pleted by the jeweller it was laid on a smooth level surface and a razor 

 blade pressed down upon it at an angle of 45° with the axis and then 

 moved parallel to itself in a direction perpendicular to the axis of the 

 cylinder which rolls under the blade, with the result that a sharp 

 well-defined spiral cut is made on the cylinder. Then the cylinder 

 is given a final polish to remove the burr caused by the cut since this 

 would interfere with the regular reflection from the surface. When 

 the cylinder hangs in the tube and the light from a straight filament 

 lamp is focussed upon it, a bright vertical line crossed by the oblique 

 cut is seen. The oscillation of the cylinder causes the cut to move 

 up and down and for this vertical motion there is no distortion. 

 The light used is a small 4-volt tungsten lamp with a straight filament 

 about one cm. long and the light is transmitted through a water cell 

 to reduce the heating effects. Except during the actual measurements 

 the windows are covered to prevent radiation. 



Originally the intention was to count the number of swings 

 necessary to reduce the amplitude from a certain value to a definite 

 fraction of that value; but this was found to be unreliable and was 

 given up. Next, photography was tried but required too much light 

 for the proper illumination of the cylinder. Finally, a transparent 

 eyepiece scale was placed in the telescope. This scale was ruled with 

 the regular divisions and was crossed by a vertical line through the 

 middle of the field cutting the divisional lines at right angles. This 

 vertical line is set on the illuminated line of the cylinder and thus 

 gives a definite point on the scale for the crossing of the oblique cut. 

 The scale is illuminated by a miniature electric-lamp placed in a side 

 tube on the eyepiece of the telescope. This is necessary as the light 

 from the cylinder is too dim to enable the observer to read the position 

 of the spiral cut easily. 



Observations of the maximum excursions on one side of the zero 

 are made and the curve of their logarithms is perfectly straight. 

 Figure 3 shows the logarithms of the maximum excursions for three 



