1819.] and on the Laws of’ the Communication of Heat. 171 
much perseverance the comparisons of all the thermometrical 
scales. 
(Note added.)—When treating of the dilatation of mercury, 
we presented a table of the results obtained by different philoso- 
phers on this important subject. The one which has been used 
in France for several years, and which is ascribed to Lavoisier 
and Laplace, is found among them. We perceived that it did 
not agree with the number which Lavoisier gives in his memoir, 
i. 310, for the apparent dilatation of mercury in glass; but we 
thought that it was the result of a subsequent and unpublished 
set of experiments. Since’drawing up our memoir, we have 
learned that these illustrious philosophers did not undertake new 
experiments on the subject ; but that an error had crept into the 
calculation of the observations ; so that the true coefficient 
deduced from their measure is ~,, instead of =}. The one 
which we found by quite a different process, >~,,, differs very 
little from theirs. This is a new proof of the accuracy of our 
observations. 
Parr [1.—Of the Laws of Cooling. 
The first views relative to the laws of the communication of 
heat are to be found in the Opuscula of Newton.* This great 
philosopher admits, @ priori, that a heated body exposed to a 
constant cooling cause, such as the uniform action of a current 
of air, ought to lose at each instant a quantity of heat propor- 
tional to the excess of its temperature above that of the ambient 
air; and that consequently its losses of heat in equal and 
successive intervals of time, ought to form a decreasing geome- 
trical progression. Kraft, and after him Richmann,} endea- 
voured to verify this law by direct experiments on the cooling of 
liquid masses. These experiments, afterwards repeated by 
different philosophers, prove, that for differences of temperature 
not exceeding 40 or 50 degrees, the law of geometrical progres- 
sion represents pretty exactly the rate of cooling of bodies. 
In a dissertation, little known, on several points of the theory 
of heat, published in 1740, and of course several years before 
Kraft and Richmann made known their researches, Martine f 
had already pointed out the inaccuracy of the preceding law, 
and had endeavoured to substitute for it another, in which the 
loss of heat increased more rapidly than by the Newtonian law. 
Erxleben § proved equally, by very accurate observations, that 
the deviation of the supposed law increases more and more as 
we consider greater differences of temperatures ; and concludes, 
that we should fall into very great errors if we extended the law 
much beyond the temperature at which it has been verified. This 
“ Newtoni Opuscula, ii. 425, + Nov. Com. Ac, Pet.i. 195, 
t Essays on Heat, p, 72. § Novi Comment, Soc, Gotting. viii. 74. 
