374 REPORT— 1885. 
to withstand the effects of earthquake motion it is necessary that the 
constructor should clearly understand the nature of earthquake motion. 
If an earthquake is regarded as a sudden blow, and we adopt rules and 
formule founded on this supposition such as are to be found in many of 
the older treatises on this subject, it seems certain that the success of our 
undertakings must inevitably be attended with uncertainty. 
Earth Tremors and Earth Pulsations. 
From time to time reference has been made in reports to this As- 
sociation to the various methods employed to record earth tremors and 
earth pulsations. Many of the observations made with delicate spirit 
levels have now been plotted. The most complete set were those made: 
at the Meteorological Observatory under the direction of Mr. Arai Ikuno- 
suke. The observations were made every three hours both night and day. 
Although a special column was built for the instalment of these levels, 
and they were protected so far as possible from changes in temperatures, 
the bubbles of these instruments wandered to and fro in a manner difficult 
to explain. M. d’Abbadie, writing to me on this subject, remarks that 
two levels may be placed parallel and yet the bubbles may move in 
opposite directions. 
Notwithstanding the untrustworthiness of level observations, they 
nevertheless have given results of interest. These are as follows :— 
1. The fact that the bubbles from time to time move back and forth 
without apparent reason. Considerable changes have sometimes been 
observed before an earthquake. 
2. The greatest movement of the bubble of a level takes place during 
the colder part of the year, which is the season of earthquakes, and also: 
the season when the barometric gradient between Siberia and the Pacific 
is the steepest. 
3. The bubble of a level continues to move long after the sensible 
motion of an earthquake has ceased, enabling us to study the slow move- 
ments which bring an earthquake to a close. 
4. When the barometer is very low, as, for instance, during a typhoon, 
the bubble of a level may be distinctly seen to pulsate back and forth 
through a range of about *5 mm. 
In addition to the fact that levels are so sensitive to changes of tem- 
perature, they have, in common with all other instruments with which I 
am acquainted, the objections that their changes between the times of 
observation are unknown. For a long time I experimented to obtain an 
instrament which would give an automatic record of earth tremors and 
earth pulsations. After many failures I think that I have at last 
succeeded in obtaining such an instrument. It is simple, cheap, and ex- 
ceedingly easy to manipulate. M. d’Abbadie tells me that it has many 
points in common with an instrument employed by M. Boquet de la Grye. 
It is briefly as follows. From a circular cast-iron bed plate resting on 
three levelling screws, there rises a tripod of angle iron about 5 feet high. 
From the top of this hangs a pendulum, consisting of a thin iron wire 
and a heavy bob. At the base of this bob, w, there is a small projection 
C (see fig. 2). As the bob and the projection were turned in a lathe, the 
extremity of c, which is flat, the centre of figure of the bob and the point 
where the supporting wire is attached are in one vertical line. Below this 
bob there is an indicating pointer, a, b, ¢, d, the full length of which is 
¥ 
