of measuring Time at Sea. 135° 
for then, its attraction being much ‘stronger than that which 
the ambient air opposes to the quitting of the particles of 
ealoric, it coals with so much the more difficulty. I have 
found, indeed, that after having heated our rods and replaced 
them again in the temperature 0, whence they had. been 
taken, it required sometimes almost twelve hours to reduce 
them to the length they were before ; that is to say, for the 
overplus of caloric totaily to abandon the interior. After 
having found this effect, and paid the greatest attention in 
my experiments, I happened at last to find in their results. 
the exactness which | had vainly sought for before; and I 
found that the glass and metals in their contractions and 
expansions followed precisely (as well as the augmentations 
and losses of elasticity of springs) the proportion ofthe de- 
grees described by the spirit-of-wine thermometer, 
These methods of proving the various contractions and 
dilatations of metals appear to me very exact: for, 1st, the 
cabinet where the instrument was placed being defended 
from the external air, no considerable change could happen 
to the wall (which was hung with tapestry, and to which 
our potences were fixed) between one experiment and the 
other: 2dly, when it does happen (and in effect it is some- 
times seen very evidently that the wall of which I have 
spoken is dilated by heat as much as the stecl very nearly), 
then, I say, this effect is announced to us by the spirit-of- 
‘wine thermometer on the one part, and by the tube of glass 
on) the other, which was very little dilatable; and which’, 
tube may, hesides, be kept in the same temperature. 
The sheaths with which our rods. were covered enabled 
them to be removed from one place to another; that is to 
say, from a stove, or from a cool place in the cabinet of. 
trial, and to adjust them on the instrument before they bad 
undergone any alteration ia their dimensions and in their 
degree of cold or heat, which would not appear practicable 
otherwise, Being well assured of these facts, I turned my 
attention to compensating the effects of heat and cold in 
my machine. 
The first idea that occurred to me, as to many others, 
was, to apply to the revulating spring a metallic thermo- 
[4 . meter, 
- 
4 
