1819.] and onthe Laws of the Communication of Heat. 165. 
of boiling water. We see that, reversing the question, this comes 
to ascertaining whether the capacities of these liquids remain 
constant when we measure the temperatures by a mercurial 
thermometer. The results of these two philosophers are different. 
According to the former, there is a slight variation m the capa- 
eity of water in the interval of the first 100°. The second, on 
the contrary, admits that the capacities are constant. This 
discordance proves, that within the limits in which the experi- 
ments were made, the variation of the capacity of bodies, if it 
exists, is very small. But such experiments are by far too 
limited to warrant us to conclude with Crawford, that the same 
principle extends to all temperatures. 
Mr. Dalton, who has considered the question, in his ingenious 
work that we have already quoted, affirms that the capacity of 
the same mass of matter does not remain constant, because a 
part of the heat is expended in producing the dilatation ; but 
that it remains invariable, if we consider the same volume. 
This assertion of Dalton is not founded on any direct experi- 
ment, and may be considered as a simple conjecture, which is in 
unison with his other ideas about the measure of temperatures. 
We shall discuss immediately the principles upon which the 
whole of his theory rests. 
It is obvious that this problem cannot be solved without 
embracing a much greater extent of thermometric scale than has 
been hitherto done. Accordingly, the experiments which we 
are about to give an account of were all made on an interval of 
300° or even 350°. 
The season of the year in which we have been obliged to 
devote ourselves to these researches not permitting us to employ 
conveniently the melting of ice, we have always employed the 
method of mixtures; but with all the precautions requisite to 
insure their accuracy. 
The bodies whose capacity we have determined, required to 
be taken from among the metals most difficult of fusion. The 
homogeneity and the more perfect conductibility of these sub- 
stances rendered them more proper than any other for the object 
which we had in view.* 
One of the greatest difficulties to which this kind of experi- 
~ ment is subject is the exact determination of the temperatures. 
We have always employed boiling water to get the capacity 
below that term; and when the nature of the bodies allowed us 
to plunge them in boiling mercury, we made use of this second 
term, as fixed as the first, and which had been determined with 
the greatest care, as we have before stated. 
But when the substance was soluble in mercury, we heated it 
in an oil bath, which, from the way in which our apparatus was 
* Each of the metals got the form of a flat plate, in order to present the more 
varface, ‘Phesedifferent plates weighed from one te three kilograinmes. 
