162 Dulong and Petit onthe Measure of Temperatures, (March, 
at the other was terminated by a capillary tube, whose capacity 
was an insignificant fraction of that of the principal tube. 
The whole apparatus being filled with mercury, and carefully 
freed from air and moisture by repeated boilings, we determine 
the weight of the mercury which was driven out, when the 
temperature was raised from freezing to boiling water. We 
shall be able to appreciate the accuracy of this process, if we 
remark that the portion of the mass which does not participate 
in the heating is insensible, and that the horizontal position of 
the tube allows us, it the case of boiling water, to apply to its 
temperature the correction depending upon the barometrical 
pressure. 
This experiment, repeated five times on different quantities, 
furnished for the dilatation sought numbers almost identical, the 
mean of which is given below. Nor have we found any appre- 
ciable difference between the effects observed in tubes of 
ordinary glass obtained from different manufactories, whatever 
was their calibre or their thickness. 
The values of the apparent dilatation at 200° and at 300°, 
have been deduced from the preceding comparison made of the 
scales of the mercurial and air thermometers. 
TABLE Iii. 
Temperatuses deduced | 4.45 apparent dilatations of|Absolute dilatations of glass sd bn apse a. 
anes mercury in glass. in vol posed uniform). P 
~ 
100° S450 38 7 00 100-°0° 
200 ste eee 213°2 
300 sore —— 352:9 
The first two columns of this table require no explanation. 
We perceive in them the apparent dilatation of mercury in glass, 
between 0° and 100°, a little less than that of MM. Lavoisier 
and Laplace, who appear to have adopted ,.,,. We expected 
a difference on the side in which it has taken place ; for in the 
work in which this last determination is stated, the authors have 
taken care to inform us that they suspected it to be too great, 
because they did not boil the mercury in the vessel which they 
employed. The absolute dilatation of mercuty which they 
deduced from it, and which has been generally employed since, 
ought, therefore, to be too great also, and this is confirmed by 
the result contained in Table II. The third column gives the 
dilatations of glass obtained by the method indicated above. 
This dilatation is increasing ; but between 0° and 100°, we find 
it as stated by Lavoisier and Laplace. ‘The last column con- 
tains the degrees which would be indicated by a thermometer 
formed of a glass plate, whose increase in length would serve as 
a measure of temperatures. We see by the deviation which has 
