168 Dulong and Petit onthe Measure of Temperatures, [MArcn, 
of the capacities which we have observed, depends, at least in 
part, upon them, is this, that the metals whose expansion fol- 
lows the most rapid law, are at the same time those whose 
capacity undergoes the greatest variations. But this question 
can only be decided by observations, which should embrace a 
greater interval of temperature than that which existed in our 
experiments. We hope soon to be able‘to throw light on this 
doubtful question. ' 
We have shown, in speaking of the dilatation of solids, that if 
we construct thermometers with the most infusible metals, and 
suppose them graduated as usual by the fixed points of freezing ~ 
and boiling water, the temperatures indicated by each of these 
thermometers would be very different., The same discordance 
ought evidently to be observed from what precedes, when we 
estimate the temperatures, as some philosophers have proposed, 
by the ratios of the quantities of heat which the same body gives 
out in cooling to a determinate temperature. In fact, in order 
that this calculation should be exact, it would be necessary that 
the body in cooling, for example, from 300° to 0° should give out 
three times as much heat as in cooling from 100° to 0°. But it 
will give out more than three times as much, because the capa- 
«ities are increasing. We should, therefore, find too high a 
temperature. We exhibit in the following table the temperatures 
that would be deduced by employing the different metals con- 
tained in the preceding table. We must suppose that they have 
been all placed in the same liquid bath at 300°, measured by an 
air thermometer. 
TGA eee ccm ses pedseuoaee. 
BPCTOULY. s a.0 sve sia c.: eee DIOR 
PMO. ste os iene ss rise Ce 
AMEMOTY.. se ene ces tw. Oeee 
TCR a ie gente ae ts eee 
Copper’. -S.: CaO ese OT 
Pitieaain sMee vec cte cnacas LT 
SA HEE AE se ge A ae 
General Reflections and Conclusion. 
Now that we have established the correspondence of all the 
thermometric scales in an extent of more than 300°, we are in a 
state to judge of the accuracy of the laws which Mr. Dalton 
conceives he has established, by measuring the temperatures 
upon a particular scale. 
The way in which Mr. Dalton has presented the principle on 
which the formation of his scale depends, does not allow us to 
consider it as any thing else than an hypothesis, which would 
have the advantage of connecting together a great number of 
phenomena by very simple relations. By substituting the indi- 
cations of that scale for those of the ordinary thermometer, we 
find, according to this philosopher : 
7 
