ON THE EAUTHQUAKK AM) VOLCANIC PIIEXOMEXA 01'' JAPAX. 129 



(e) Periodic movements of tieveral dai/s' duration, and wandering 



vf the pendtdums. 



In order to determine the existence or non-existence of periodical 

 movements greater than twenty-four hours, the mid-day position of 

 instruments A and J were plotted on squared paper. This was done 

 for dates between October 13 and December 4, 1894. Because the 

 instrument at K during the period wandered so rapidly towards the east, 



Fig. 4. 



and had in consequence to be repeatedly re-adjusted, its records have not 

 been considered. Whether the mid-day reading is the best one to 

 represent the mean daily position of the instruments must for the present 

 be left until a second examination of the diagrams has been made, when 

 the analysis may be continued up to the end of the first week in 

 February 1895 (fig. 4). 



FiG. o. 



fS6>4 



i7 2f ZJ ZSMkZ 



/4 JS 22 26 30 



JBer. 



By drawing a free curve through the diagrams of mid-day positions 

 of A and J, and reducing this horizontally and vertically one-fourth, tlie 

 two curves shown in the accompanying figure are obtained (fig. 5). 



During the complete period considered neither of the pendulums 

 has changed much in its mean position 

 1895. 



Whatever slight changes have 



