of measuring Time at Sea. 201 
greatest degree of heat that can happen to it in different cli- 
mates ; without which (the metal being by this heat a little 
annealed, and metals forged more or less and then annealed, 
dilating unequally, and besides, acquiring more extent by this 
annealing) the machine would gain after having experienced 
great degrees of heat; the parts of the metal, however little 
they may be annealed, do not return to the first contraction 
which had been given them by tempering, forging, &c.; 
and the same may be said of other parts of the regulator. 
Observation VI. 
On the size of the machine, 
I think it ought not to exceed much the size of a varia~ 
tion compass, of which we can never complain. 
This watch in itself is of a sufficiently small size; but it 
appeared to me, that to give greater security to all the parts, 
to put them more out of the way of rust, of excess of heat 
of short duration, and other accidents which may happen in 
a ship, it would require a case; and it would be right that 
all the parts of suspension, as well as the watch, should be 
shut up in a strong box with a lock. Besides, to diminish 
the effect of shocks, it would be necessary, as I have ex- 
plained Article VIII. Part III., that it should be stopped in 
its motions by the cushions, which can hardly be adapted 
but to the sides of such a box. 
Recapitulation. 
After having explained the principles which have directed 
me in the different parts of my work, being the fruit of 
twenty years researches and labour, I shall attempt to show 
that by means of it we shall have the best measure of time 
at sea. 
The following is the way I prove it: 
It is evident by the first part, that this measure of time must 
consist in the most advantageous application of the balance 
and its spiral spring to the clock; that is to say, in a watch 
perfected as much as possible. To prove, therefore, that we 
are arrived at the best measure of time at sea, it is necessary 
to show that we have used the best methods to perfect the 
watch, that is to say, to correct its defects, which consist, 
3 as 
