18 Proceedings of Royal Society of Edinburgh. [sess. 
intensity to about *00004 C.G.S. units, and the action appears 
therefore more magnetic than mechanical. At 5J minutes past 
four the vibrations pass the limit of photographic action, and this 
perhaps marks the time of commencement of the more intense oscil- 
lations. The trace appears faintly at about 19 J minutes past the 
hour, less faintly at the extreme ends of the amplitudes than at 
the middle of the curve, and it continues faint till about 26 minutes 
past, when the trace darkens, but continues disturbed till 36 
minutes past four, when the normal curve is resumed. The time 
Fig. 2. 
of vibration of the magnet has been ascertained to be 8 seconds, 
and obviously, as this factor, together with the period of dis- 
turbance, must influence the motion of the magnet, the irregu- 
larity in the trace is perhaps to a considerable extent due to that 
cause. 
The trace shown here (fig. 3) is that of the vertical force magneto- 
graph. As pointed out elsewhere, this instrument is very sensitive 
to shocks, resulting in dislocations of the curve. No such disloca- 
tion is detected during the earthquake. The unusual thickness of 
the vibration trace shows that the cause of the disturbance could 
