ON SEISMOLOGICAL INVESTIGATION. 209 
not necessarily a horizontal pendulum, but an ordinary pendulum, a 
balance, and perhaps even a magnetometer, shows considerable motion 
within its case, the question arises whether there be not air currents 
existing within the cases which cover such instruments. With compara- 
tively heavy horizontal pendulums in well ventilated cases in Japan 
tremors were always small and of rare occurrence, but with light 
pendulums in similar cases tremors of a pronounced character nearly 
always occurred between midnight and about six in the morning. With 
the light pendulum at Shide, beneath a fairly tight case, I found tremors, 
whether the inside of this was lined with thick felt, and the supporting 
pier covered with the same material, which kept the surroundings of the 
instrument at a fairly uniform temperature, or whether such coverings 
were removed. Covering those portions of the column which were inside 
the case with cement, and painting the surface of the same, did not 
destroy the intruders. Another experiment was to replace the large 
doors of the case with fine gauze, thus giving the instrument considerable 
ventilation ; but, as will be seen from the records (November 21-30), no 
great improvement was effected. By means of a very fine column 
of smoke from the spark at the end of a thin joss-stick, joints in the 
covering cases were tested for draughts. The column of smoke was also 
placed before a small hole usually closed by a cork, to see if there was any 
tendency in the air to enter or come out from the case, but no indication 
of the same was obtained. 
One very marked observation was that a strong tremor storm would 
suddenly cease, or be at least greatly altered in its intensity, by opening 
the door of the case for one or two minutes.! 
Although a sudden change of this last description has occurred with- 
out opening the doors, we have in this observation an indication that 
by some means or other, which do not seem to be effects due to differ- 
ences in temperature in different parts of a case, air-currents are from 
time to time established within a case, the mechanical working of which 
can be more or less destroyed by simply opening the door of the case. 
One cause of such currents may be due to the different rates at 
which aqueous vapour is absorbed or given off at different points within 
the covering, and if these are steady they may set up a steady set of long 
period displacements in a light pendulum. 
By introducing a tray of calcium chloride inside the case, violent 
“movements have resulted, which only ceased after the desiccating agent 
‘was removed. 
_ These facts, coupled with the fact that tremors were apparently 
greatly reduced by surrounding the boom with a trough or wind-guard on 
- three of its sides, lead to the conclusion that air-currents are from time 
to time generated within casings such as I have employed, which result 
in movements which are with difficulty separable from those which are 
attributed to motion of the supporting pier. 
The fact that tremors occur when there is a slight fall in temperature 
outside the case, whilst the fall inside the same would be comparatively 
small, suggests the idea that at such times, although they have failed 
detection, there may be streams of air passing through the joints of the 
coverings. The unlikelihood of this is, however, referred to in the next 
section. 
' In some instances, however, the opening of the door seems to have brought 
. # tremor storm into existence. 
1896. | P 
