1830 ] 
by Messrs. Dulong and Petit. 
279 
Mr. Dalton, who has touched on this subject, in the ingenious work we have 
before cited, supposes that the capacity of the same mass cannot remain constant, 
because part of the heat is employed in effecting expansion ; but that if we confine 
ourselves to the consideration of the same volume, it will be constant. 
This doctrine of Mr. Dalton’s is not founded on any direct experiment, and may 
be viewed, in fact, as a mere conjecture connected with his other ideas relative to 
the measure of temperature, and to which we shall return when we have discussed 
the principles on which his whole theory is based. 
We may, in the mean time, apply to the present question the same argument we 
used with relation to expansion ; viz. that we cannot expect to resolve the problem, 
except by taking into our consideration a much more considerable portion of the 
thermometric scale than has yet been done. Accordingly, the experiments we 
are now about to give an account of, have all been made in an interval of 300° and 
even 350°. 
The season during which this part of our experiments has been conducted not 
allowing us conveniently to use the tempei’ature of melting ice, we have confined 
ourselves to the method of mixtures, but with every precaution necessary to ensure 
the utmost precision. 
The substances, the capacities of which we have determined, were necessarily 
chosen amongst the least fusible of the metals. The homogeneity and perfect 
conductibility of these bodies render them fitter to be the objects of such ex- 
periments as we bad undertaken, than any other 2 . 
One of the greatest difficulties which attends this kind of inquiry, is the exact 
determination of the temperatures. We have always used boiling water in our ex- 
periments, to measure the capacity below that point ; and for temperatures above 
it, boiling mercury, when the nature of the body would permit. We have used 
this point, considering it to be equally fixed with the other, and as having determin- 
ed it with the greatest care, as we before noticed. If however the substance was 
soluble in mercury, we heated it in an oil bath, which, by the arrangement of the 
apparatus, could be kept steadily at the same temperature for about a quarter of 
an hour. Lastly, to avoid the error which might have been occasioned by an in- 
equality of temperature in different parts of the mass — the liquid was stirred con- 
tinually at the period of attaining the maximum temperature ; and, by a thermome- 
ter with a constant volume, the mean temperature, which was of course that of the 
body, was ascertained. Fixed oils, it is known, acquire a much higher degree of 
fluidity when heated — the film, therefore, which adheres to bodies plunged into 
them, is very thin : we have not, however, neglected to allow for the heat deriva- 
ble from this small addition of matter ; although in a majority of cases this correc- 
tion has only had reference to very minute quantities 3 . 
When the body under experiment had been heated to a certain temperature, 
(which was measured in the way we have just described,) it was plunged, as quick- 
ly as possible, into a large mass of water, and the rise of temperature observed of 
this liquid, as soon as an equilibrium, was established. It is in the determination of 
this element of the inquiry that the greatest possible precision is required, in order 
to obtain results that can be depended on. Our practice has been, to employ a quan- 
tity of water so large, that the rise of temperature could never exceed 5 or 6 centi- 
grade degrees ; and to measure this, a thermometer was used, with a scale, such, that 
the one hundredth part of a degree could be accurately read off. The water was 
contained in a vessel of very thin tinned plate, insulated upon a support which 
touched only at three points. This vessel, in each case, received part of the heat ; 
but as its weight and specific heat were known exactly, we could, of course, in each 
case, allow for the effect it produced. 
In the greater number of experiments, the water was cooled down previously ; so 
that when the heated body had imparted its caloric, the whole should be of the tem- 
perature of the surrounding medium : in others the beating process commenced 
from this latter temperature. The first method has appeared to us more exact • it 
requires no correction : for the water, immediately after the heated body is plung- 
ed into it, acquiring a temperature but little removed from that which must finally 
2 Inorder to increase the surface as much as possible, the pieces with which we ex- 
perimented were formed into the shape of very flat rings. They weighed from one to 
three kilogrammes. 
3 This correction was made to depend upon the weight of oil which adhered to the 
fing. To determine ibis, we made in each case a preliminary experiment, by weigh- 
ing the ring on withdrawing it from the oil bath. At 300® C. the increase^ of 
weight never exceeded 3 to 4 decigrammes. 
