250 Dulong and Petit onthe Measure of Temperatures, [APRiL, 
observation. In considering the question under this point of 
view, it would be necessary to know, in the first place, the law 
of the capacities for a certain body, and to determine then, by 
direct observations, the quantities of heat lost by the same body 
at certain fixed terms of temperature indicated by the air ther- 
mometer. Then by multiplying the velocities of cooling deduced 
from the preceding law by the corresponding capacities, we 
should obtain the absolute losses of heat. Itis notin the interval 
of the first two or three hundred degrees of the centigrade scale 
that we can hope to verify the accuracy of these consequences. 
The variation of the capacities not beginning to become very 
sensible till we pass that term, it would be necessary to observe 
at temperatures of 5 or 600°. It is easy to see the difficulty of 
such a kind of observation. However, we have succeeded in 
constructing apparatus fit for the purpose ; and we have already 
made a great many observations relative to this subject. But as 
our results do not yet present all the regularity which we expect 
to be able to give them, we have determined to delay their pub- 
lication; and so much the more willingly, that the question 
which it is their object to answer does not come within the limits 
of the prize proposed by the Academy. 
The method which Mr. Leslie employed for measuring the 
radiating powers of different surfaces is very good for making us 
acquainted with the radiating heat lost by a body at all temper- 
atures. It is well known that his method consists in estimating 
the radiation of a body by the heat communicated to an air or 
mercurial thermometer placed at a certain distance from the hot 
body ; and to render the effects more sensible, this thermometer 
is placed at the focus of the reflector. 
It was by means of this apparatus that Laroche obtained the 
result which we have mentioned above. Among the series of 
observations made by this method, there is one which extends 
indeed to very high temperatures ; but it cannot be of any utility, 
because the temperatures were determined by a process founded 
on» the supposition that the capacities were constant. The 
numbers which represent the losses of heat are besides affected 
by another error, proceeding from the heat of the focal thermo- 
meter being too great, because the inaccuracy of the Newto- 
nian law had already become very sensible. But to show that 
our law satisfied the observations made by this process when they 
are freed from the causes of error of which we have just spoken, 
we shall apply them to the series given in the same memoir 
which do not go beyond the limits in which the variation of 
capacity produces no sensible effect. These series are the 
radiation of aniron crucible filled with mercury. Here the tem- 
perature of the body not having exceeded 200°, we may suppose 
the specific heat constant. We may likewise neglect the cor- 
rection which the mercurial thermometer would require to bring 
it to the airthermometer ; because this correction is very small, 
Fe ; 
= 
