of measuring Time at Sea. 197 
veniences. It was at first difficult to be assured that the spring 
was sufficiently straight, and that the balance was attached in 
such a manner, that its weight acted in a line along the middle 
of the breadth of the spring throughout its whole length : 
without this, however, it produced a very disadvantageous 
friction, and a difficulty in each vibration. Moreover, the 
weight of the balance was not sufficient to stretch this spring 
perfectly ; it was hardly possible for it not to be bent a little 
in some part of its length: these curvatures diminishing by 
shocks and heat, or augmenting by cold, there arose irre- 
gularities difficult to prevent. Lastly, the distance at which 
the elastic force acted, being only equal to half the width of 
the spring, the least differences which might happen in the 
situation of this spring, whether by the small play of the 
balance in its holes, or by other causes, could not but have 
some influence on the manner in which it acted; this does. 
not happen in the spiral spring, which acts always at a con- 
siderable distance from the axis of the balance. 
All these inconveniences are prevented by the harpsichord 
wire; itis so small that it can have but little influence on 
the vibrations ; it may be made much shorter ; it is exactly 
stretched, and without curvature throughout its whole length ; 
and being round, we may be certain, at first sight, that all 
its parts agree with the axis of the balance. 
To this it may also be added, that, by the operation of 
drawing the wire, we are assured that the substance of which 
it is composed is homogeneous and pliant, such as it ought 
to be for this suspension. 
I have said it is necessary for the wire to be fine: expe- 
rience has proved to me that without this it would require 
to be very long, which would render the machine cumber- 
some. Having taken to suspend a balance a thicker harp- 
sichord wire, of about four inches long, I remarked that the 
motion of the regulator lost with the greatest readiness until 
it was reduced to describe only four or five degrees, and 
then it remained as long a time in motion as if it had been 
either very long or very small; whence I concluded that the 
motion is not lost so readily in a large are, that because the 
parts touched they formed an obstacle in some degree insur- 
N3 mountable, 
