of measuring Time at Sea. 63 
obscurity on this capital subject, I think it will be best to 
examine it a little more in detail. 
I acknowledge that, in watches, where the very confined 
space does not permit us to apply all the resources of the 
art, it would be very dificult to determine, whether some 
slight differences in the times of the long and. short vibra- 
tions do not sometimes, as well as slight frictions, produce 
compensations, whence results a greater degree of accuracy, 
or rather less inequalities; but in the present case, where, 
being master of space, we may use methods less uncertain, 
it may be demonstrated, and I shall prove, that the isochro- 
nism and perfect freedom of the regulator are the only means 
to obtain a greater degree of truth. 
In effect, friction is a thing subject to a thousand varie- 
ties incompatible with much exactness. Let the regulator 
in its vibrations experience some friction; it necessary fol- 
lows, that the quantity of this friction will vary according 
as the contact of the air alters the polish of the rubbing sur- 
faces, according as they alter each other, and the softest 
body leaves its parts on the hardest, according as the oil 
which we apply to soften the friction becomes more or less 
fluid, &c. 
_ Let us suppose, also, that the vibrations of the regulator, 
abandoned to itself, are not isochronous, but that by some 
mechanical artifice we happen to render them all of equal 
duration ; (by friction, for example, or, as this operates in 
some clocks, by the curves of the anchor escapement, or by 
that with a double lever, &c.;) I maintain, that such an iso- 
chronism, being subject to a thousand uncertainties, can 
never give the necessary precision in a time-keeper at sea. 
In effect, besides the varieties of friction, M. Le Roy* has 
demonstrated, in treating of the pendulum, that a diminu- 
tion of the arc of vibration arising from that of the motive 
force, from the clogging of the wheels, or from that of 
the regulator, requires in each of these cases, in order to 
be compensated, the curves of the anchor to be altoge- 
ther different; and likewise longer or shorter pallets in 
the escapement with a double lever. Now what he has 
* * See my Memoir on Clockwork, &¢. published in 1750, y 
, said 
