Prof. Cavalleri on a New Seismometer. lll 
ratcheted in such a way that it can easily go up, but cannot 
descend, as seen in the figure (Plate I.). Thus the index will 
mark the elevation of the ground or vertical undulations, the 
ratio between the two arms of the iever being taken into account. 
Instead of the lever we may use a cylinder of cork a little larger 
than the weight, which cylinder must run easily between two 
fixed vertical side-pieces or guides, as represented in the figure. 
As far as the cork, by the push of the weight, has been lowered 
it will remain fixed there, being light and held by its own elas- 
ticity so fast that it is impossible for a shock alone to move it 
from its position. An upright scale, suitably divided, placed at 
the side of the cylinder, shows the height of the vertical wave,— 
as it rises with the ground, while the cylinder is depressed. 
This last method is, I think, superior to that which I constructed 
with the lever. 
This instrument is so delicate, that by giving, if necessary, 
a larger diameter and a greater length to the spiral, and a cor- 
responding weight, we may succeed in marking any slight ele- 
vation of the soil performed in a relatively long period of time. 
For example, we might record an elevation with a velocity of a 
millimetre per second, or even still slower. Knowing by obser- 
vation that elevations or depressions of the soil occupy a very 
short time, I thought it useless to give the spiral a longer time 
of oscillation, It may happen that the first vertical wave is one 
of depression and not elevation ; but in this case also the instru- 
ment will accurately note the vertical movement, in consequence 
of the elevation which succeeds the depression. It may happen 
also, and in fact often does, that the weight gradually stretches 
down the spiral and falls a certain degree (by loss of elasticity, 
namely) ; but in this case the short arm of the lever is carried 
down also by the weight, and the other arm the place fixes, on 
account of the ratcheting already mentioned, so that the space 
found between the weight and the lever below it will always 
indicate the height of the vertical wave. I will not conceal an 
objection which may be alleged against this method of measuring 
vertical undulations—an objection which, with much greater rea- 
son, may be raised to other seismometers,—viz. that when vertical 
undulations are frequently repeated, the instrument will mark 
sometimes more and sometimes less than the real altitude of the 
earth-waves, according as the time of the undulations is tauto- 
chronous with that of the oscillations of the pendulum or the 
contrary. If the spiral perform its oscillations more slowly, 
this defect will generally be very small; and besides, these ver- 
tical shocks being almost always confined to a single pulse, the 
instrument will mark accurately in most instances. 
Lastly, there remain mixed shocks, 7. e. those which, besides 
