HORIZONTAL PENDULUMS. 131 



each of which generally lasts for some clays. Their effects must 

 not be overlooked in determinations of the force of gravity and in 

 other delicate measurements of geo-physics. 



The period of the pulsatory oscillations seems to depend, to 

 a slight extent, on the range of motion, the period of the larger 

 movements in a given interval of time being usually a few tenths 

 of a second longer than that of the smaller movements in the 

 same interval. Again, the general average period of the pulsatory 

 oscillations seems to remain nearly constant for a certain length 

 of time, it may be even for several days. 



The rate of motion of the smoked paper is, in the present 

 forms of my Horizontal Pendulums, too rapid to allow of 

 pulsations of several minutes period, as observed by Paschwitz, 

 Milne, and others, being discerned. 



Pronounced storms of pulsatory oscillations have been 

 generally observed, when strong winds prevailed in Tokyo or its 

 vicinity, as was the case from the 6th to the 7th, September 1898. 

 (See fig. 9). Nevertheless, these movements have been met with in 

 very quiet weathei's, as was the case on the 14th and 15th October, 

 1898. (See fig. 10). For the sake of reference, I give, in PL 

 XII, the weather maps for these days. 



In the storm of pulsatory oscillations on the 6th and 7th 

 September, 1898, the maximum range of motion was 0.2 mm. in 

 each component. The average period of the oscillations of larger 

 amplitude was 6.0 seconds, while that of the oscillations of smaller 

 amplitude was 5.6 seconds. 



In the storm of pulsatoi-y oscillations on the 14th and 15th 

 October, 1898, the maximum range of motion was 0.1 mm. in 

 each component, the general average period being 6.8 seconds. 



It must be understood that the period of the pulsatory 



