110 Prof. Cavalleri on a New Seismometer. 
must therefore substantially agree with the laws of the same 
philosopher. The third would agree if the weight of the spiral 
were nought, or might be so considered; but as one cannot 
attach to the spiral a weight which would render the weight of 
the spiral itself evanescent, since it would draw down the spiral too 
much and cause it to lose its elasticity, so the weight of the spiral 
has always a sensible effect, and tends to retard the oscillations : 
some advantage is gained by the use of spirals of tempered steel. 
The second law has no counterpart (riscontro) among those of 
Coulomb, as it depends on conditions not found in twisted 
threads, on which the laws of elasticity of torsion depend. With 
the same wire and of the same length, spirals can be formed 
which will have widely different oscillations by enlarging or nar- 
rowing the diameter of the coils. 
These laws ascertained, it is easy to find the conditions best 
adapted to our purpose. The weight attached to the spiral must 
be of a certain size in order to produce a strong reaction, and to 
move the markers which we are about to describe. The oscilla- 
tions ought to be slow, so that the time employed by the rising 
and falling of the ground may not exceed that required to trans- 
mit the motion from the top of the spiral to the weight itself. 
I have therefore given to the spiral, measured along its axis, a 
length of 80 centims., and attached to it a weight of 1:2 lalog. 
The diameter of the cylindrical spiral is 5°3 centims. It con- 
sists of ninety rings, and vibrates seconds: the diameter of the 
wire is about 3 millims. I have constructed it in the following 
manner :—A strong iron bar fixed in the wall supports one end of 
the spiral ; a cylindrical weight of equal diameter to the spiral 
is attached to the other extremity. This weight oscillates freely 
within an iron ring secured to the wall. The spiral is enclosed 
in a kind of cylinder in which it can freely oscillate vertically, 
but not horizontally. The weight terminates in a point, and 
rests on the short arm of a lever very easily moved; the other 
arm of the lever, by means of a graduated quadrant, serves as 
an index. All this apparatus of lever, quadrant, and imdex, is 
securely attached to the wall, and is united to the bar by which 
the spiral is suspended. Now let us suppose that the ground 
has been suddenly elevated by an earthquake. It is evident 
that, when the wall and the bar which holds the spiral are raised, 
the agm of the lever, being part of the same rigid common system, 
will rise also. But for a certain time the weight attached to the 
spiral will remain unmoved in its place, because a sensible time 
is necessary to communicate motion from the top of the spiral to 
the weight itself. Meantime the short arm of the lever will be 
pushed by the weight, and being very light, it will fall, while 
the longer arm will rise and record the elevation. This arm is 
