108 Prof. Cavalleri on a New Seismometer. 
be impossible to make them out accurately. However, if we can- 
not always accurately tell the direction of the waves which have 
different horizontal inclinations, we can at least discover the 
direction of the principal one. Since the oscillations of the pen- 
dulum, especially if it be a long one, must be very small in com- 
parison to the motion of the terrestrial crust lying beneath it, 
we hope no argument will be drawn thence against increasing 
the sphere of application of our seismometer. Another desi- 
deratum is to mark the time when the earthquake commences. 
In some seismometers which I saw at Bologna, and in others of 
which I have read, at the moment the shock occurs, the pendulum 
liberates a rod or a weight which in some manner (several me- 
thods are employed) stops the motion of a timepiece which is 
placed near. In others, a pencil moved by clockwork draws a line 
on acard divided into twenty-four parts, according to the hours of 
the day; an irregularity in the line will prove the occurrence of an 
earthquake. This manner of marking time has the advantage of 
compelling the observer to take daily note whether an earthquake 
has occurred or not; but is attended with the enormous incon- 
venience of requiring a person to attend to the instrument daily, 
perhaps for years, before the occurrence of the desired phzeno- 
menon. In our apparatus it is just the reverse. The timepiece, 
constructed with a main spring and a strong balance, and secured 
to the wall, is always wound up, but does not go. The instant 
a shock moves the pendulum, however slightly, a lever which 
retains the balance in a position favourable to its easy disengage- 
ment, is set at liberty, and the timepiece begins to mark time. 
The index is placed at zero, and can mark twenty-four hours, the 
dial being divided into twenty-four parts. Itisevident that at what- 
ever hour of the twenty-four a person perceives the motion of the 
clock or the displacement of the lever, or hears the ticking of the 
timepiece, &c., he can accurately tell at what hour the earthquake 
took place, by subtracting from the true time, as given by chro- 
nometers, the hours recorded by that which is attached to the 
seismometer. A whole day could not possibly elapse without 
the attention of some individual being attracted to the instru- 
ment, especially if (secured in a glass case) it were erected in a 
frequented and easily accessible place. 
But seismometers ought to mark not only undulatory or hori- 
zontal motions, but subsultatory or vertical also, as well as mixed 
ones. With regard to vertical upheavals, I have made use of a 
property which T observed belonged to spirals or elastie coils 
(cliche), viz. their power of vertical oscillation. I reflected that 
as the oscillations of the pendulum mark in a horizontal diree- 
tion the horizontal undulations of the ground, so the vertical 
oscillations of the spiral might mark the vertical elevations of 
