461 



examination of these curves it appeared, as had been antici- 

 pated, that the maxima and minima of the irregular movements 

 usually succeeded each other with very great rapidity, their 

 interval being, on the average, about 40 seconds, and theit 

 magnitude varying in these observations from 10^' to 60''. 



The observations were resumed on the 23rd of October, 

 and continued to the 26th, two series of an hour's duration 

 being made each day, in the hope of detecting some law go- 

 verning the movements. No such law however could be traced, 

 nor did there appear to be any connexion between the curves 

 representing the march of the changes, at different hours of 

 the same day, or at the same hour on successive days. 



The observed variations in these observations being small, 

 it was suggested by an experienced friend, to whom Prof. 

 Lloyd showed the results, that they were not true magnetic 

 changes, but merely the errors of observation incidental to 

 vibratory movement. In order to test this supposition, an 

 unmagnetic bar (the brass detorsion bar of Gauss's appa- 

 ratus) was substituted for the magnet, to which it cor- 

 responded in dimension ; and being provided with a mirror, 

 was suspended by two parallel threads, and made to vibrate. 

 The time of vibration was adjusted, by varying the interval 

 of the threads, so as to differ little from that of the magnet. 

 The successive elongations were then observed, as in the 

 case of the magnet, and the means of each pair taken. The- 

 variations in these means (which could arise from errors of 

 observation or mechanical changes only) bore no com- 

 parison whatever in magnitude to the corresponding varia- 

 tions of position of the magnetic bar, thus showing that the 

 latter were truly the results of the operation of magnetic 

 forces. 



The rapidity with which these changes thus appeared to 

 follow each other, held out the hope that they might be em- 

 ployed in determining differences of longitude ; and it only 

 remained to ascertain, for that purpose, whether variations so 



