118 Dulong and Petit on the Measure of Temperatures, [Frs. 
ment of the expansion of solids, even below the temperature 
of hoiling water, we are terrified at the much more nume- 
rous obstacles which would accompany the same determina- 
tion at elevated temperatures. After a careful consideration 
of all the experimental resources which we could hope for, the 
uncertainty of success and the enormous complication of 
apparatus ‘that would have been required, determined us to 
undertake the direct measurement of the absolute dilatation of 
mercury. This is the object of the following chapter. 
Of the Absolute Dilatation of Mercury. 
The knowledge of the absolute dilatation of mercury became 
essential as soon as it was perceived that heights might be 
exactly measured by ineans of the barometer. Nor is the datum 
less useful in many physical experiments. Accordingly, there 
are few determinations which have given rise to so many 
researches. But notwithstanding all the precautions of experi- 
menters to obtain accuracy, there are few examples of greater 
discordance than are to be seen in the results which they have 
obtained. The following are some of them, 
Absolute Dilatations of Mercury. 
Dalton. ...... niaathaetimadisa sz Lavoisierand Laplace .... +4 
Lord Charles Cavendish .. 4 Haellstrom. ............ aie 
1 EL Caen Ribak hes geek li sz Lalande and Delisle...... <4 
Wermenal TRO is pains gis ye yim aie CURNOIRG a0 op eatbehs 2 ogee 
SOGeEDUTON. .... ss eon ee oe sy 
The greater number of these determinations haye been calcu- 
lated by adding to the apparent dilatation of mercury in glass 
the proper dilatation of this last substance. And as we were 
long uncertain of the true expansion of glass, the preceding 
results behoved to share this uncertainty. 
Deluc, Casbois, and General Roy, endeavoured to measure 
directly the real dilatation of mercury by the increase of the 
barometrical column, occasioned by a known variation of tem- 
perature. The results cbtained in this way are very inexact. It 
would be easy to assign the reason by discussing the methods 
employed by each of the three philosophers whom we have 
named. But it would be requisite to enter into details which 
might become tiresome. Besides, the experiments to which we 
allude apply only to temperatures below 100°, and it is beyond 
that point that we require particularly to know the real dilatation. 
of mercury. It became of course necessary to have recourse to 
new methods. The one which we shall now describe appears to 
us susceptible of all the precision that can be desired. 
It is founded on this incontestible law of hydrostatics, that when 
two columns of a liquid communicate by means of a lateral tube, 
the vertical heights of these two. columns are precisely the 
inverse of their densities. If then we could measure exactly 
