Disturbance of the Needle during an Aurora. 



125 



effect of the diurnal variation, may be formed by the comparisons con- 

 tained in the annexed table, in which the observations made on the 

 10th of July, are compared with the mean of those made on the same 

 hours, two days previous, and two days subsequent to the 10th. 



In this table the columns are arranged as in the foregoing one, ad- 

 ditional columns being introduced to show the number of observa- 

 tions from which the mean is obtained, and to give the differences 

 between the changes of the 10th and the mean changes. The com- 



