560 REGNAULT ON THE 
Commission of Steam-Engines, with a liberality that will be ap- 
preciated by all friends of science, placed the necessary funds for 
the execution of the work at my disposal. I propose to publish 
in a series of memoirs the results at which I shall arrive in these 
researches, without confining myself each time to any particular 
order, and reserving for a later period the summary of the whole, 
when the researches shall have become sufficiently advanced. 
In this first memoir I shall treat of the elastic force of aqueous 
vapour at different temperatures. 
The elastic forces of aqueous vapour at different temperatures 
have occupied the attention of so many natural philosophers, 
that one would suppose very slight discrepancies must exist in 
their numerical values. This however is not the case, as may 
easily be seen by comparing the results which have been pub- 
lished by different experimenters ; very great differences are ob- 
served, even within the limits of atmospheric temperatures, and 
it is impossible to decide which of these numbers ought to have 
the preference. I shall not stop to retrace the history of the 
researches made upon this subject; the most important are 
known to all natural philosophers. I pass therefore immediately 
to the description of my own experiments. 
To establish with any degree of precision physical data, it 
does not suffice that a single mode of experimenting be followed. 
If the results obtained by one method are not in accordance 
with the facts which already exist in the science, it becomes 
often difficult to decide which of these ought to be preferred ; 
for directing our choice in this case we can only be guided by 
the more or less favourable opinion which we can form of the 
mode of operating, and the amount of confidence with which 
the skill of the operator inspires us. It is necessary to repeat 
the experiments by a variety of methods, and to employ even 
the modes already adopted by philosophers who have been en- 
gaged in the same pursuits, at least when these modes are not 
absolutely defective. It is necessary to show that all these 
modes of procedure, when properly carried out, lead to the same 
result; or if they do not, we must point out by direct experi- 
ment the causes of error in the defective methods. 
This proceeding is necessarily long and tedious, but it ap- 
pears to me the only one calculated to introduce certain nume- 
rical data into the science, which ulterior experiments can but 
modify in an insignificant degree. I endeavoured to follow this 
