Observations zuitli Kater's Pefidulums. 17 



right angle, and carrying a steel spike d ; ff i^ anotlier brass 

 plate with a recess cut out at one end to allow the arm be to 

 pass ; e, e, are screws with conical ends, passing through j^ and 

 bearing into holes in the line of bending of the plate abc, thus 

 forming pivots about which it can turn easily ; ^ is a circle 

 traced on the upper side of ab, its centre being at the same 

 distance from the line of the pivots as is the point of d. When 

 in use the point of the spike d rested against the south window 

 of the cylinder, the arm ab being just lifted off the plate _^ 

 by the pressure. A mass /^ being placed on g, the lever abc 

 transmits a horizontal thrust to the cylinder equal to the weight 

 of k. The plate _^ was screwed to a wooden block, in its turn 

 screwed to the top of an iron drum filled with water. 



With a brass 1 oz. weight the following readings were obtained- 

 Loaded. Unloaded. 



6-2 ... — 



— ... 7-0 

 6-2 ... — . 



— ... 7-0 



Consequently a horizontal pressure in the plane of oscillation 

 of the pendulums equal to the weight of ^ oz. would produce a 

 deflection of O'l scale divisions. 



(c) To test the effect of small impulses. 



Taps with the finger dealt as lightly as possible to the 

 south window caused deflections of several scale divisions ; the 

 resulting oscillations were however damped out after two or 

 three swings. 



The sensitiveness and delicacy of the apparatus being thus 

 demonstrated, the pendulum (No. 11) was set in vibration 

 through an arc about equal to the largest employed in the 

 coincidence observations, and examined at intervals during a 

 quarter of an hour. On no occasion could the slightest trace of 

 oscillation be detected, the image of the scale remaining to all 

 appearance absolutely steady on the crosswire. 



The distance betweeen the knife-edges and the plane of 

 support of the cylinder was about six inches ; and since the 

 angular excursion, if any existed, could not as shown exceed 

 0"*05 on either side of the vertical, the maximum possible linear 

 displacement of the knife-edge from its equilibrium position may 

 be put down as "00003 inches, which is about one ten- 



