Intelligence and Miscellaneous Articles. 



425 



the wall, the tendency to elliptic motion was greatly diminished, and 

 other sources of error entirely got rid of. 



• In experimenting with this pendulum, I have again confined my- 

 self to very small arcs of vibration. One of the many advantages of 

 this method is, that the ball and lower portion of the pendulum can 

 be easily enclosed, and thus protected against currents of air. This 

 was at first done by placing round it a rude enclosure of boards ; but 

 afterwards a square pyramidal frame was constructed, and glazed on 

 the four sides, one of them opening as a door ; so that the ball vi- 

 brated, during the experiments, within the glass case, the top only 

 being left open to admit the wire. 



My experiments with this pendulum are divided into two classes, 

 according to their different duration. Those of the first class have 

 an average duration of about an hour each ; and the values of a and b 

 (or semi-axes major and minor), at the beginning and end of each, 

 have been observed and recorded. These are afterwards subdivided 

 into six groups, according to the value and sign of b. 



The second class contains experiments of much longer duration ; 

 two of them extending through more than 37 hours each, with only 

 a single impulse of the pendulum. The ellipticity of these has not 

 been observed ; but as the mean arcs are small and the direction of 

 the elliptic motion is twice reversed while the plane of vibration is 

 passing through 180°, the apsidal motions of the whole mass must 

 almost entirely destroy one another. It is somewhat remarkable that 

 the changes from direct to retrograde (or + to — ) motion, fall nearly 

 on the same parts of the azimuth circle here, as they did on that at 

 St. Nicholas ; the change from — to + being at about 60° or 70°, 

 and that from — to + at about 140° or 150°, in both instances. 



Experiments of short duration, arranged according to values of b. 

 i4^j^i\^'*-^!^y^Ti!,, Elliptic Motion Retrograde. 



N.B. In finding the mean values of b and a for each group, re- 

 gard has been had to the different durations of the experiments. 



