REGNAULT’S HYGROMETRICAL RESEARCHES. 653 
It seems to me impossible to bring all these circumstances 
into the theoretical calculation of the instrument, and I think it 
more prudent to employ theoretical considerations only in the 
investigation of the form of the function, and to determine after- 
wards the constants by experiments made in fixed conditions. 
This mode of operation appears to me the more necessary, as 
there is much uncertainty as to several of the numerical elements 
which enter into the calculation, especially as to the specific heat 
of the air, that of the vapour, and the heat absorbed by the water 
when it is evaporated in the air. I shall indicate, at the con- 
clusion of this memoir, processes which will admit, I think, of 
determining these elements with precision by direct experiments. 
Thus, let us take 
t—t') 
e=ap—Sk ld deen ag C4 
and let us seek whether this formula, applied to the calculation 
of the indications of a psychrometer placed in very various cir- 
cumstances, can give in all cases the real quantity of humidity, 
by determining agreeably the constants A and B. If the for- 
mula thus determined cannot represent the quantity of humidity 
which exists in all cases, we may suppose A and B functions of 
t, or of z', or of (¢ — t') that, for more simplicity, we shall take 
from the form a + O¢ or &e. 
] 
a+6? 
I began by seeking whether the temperature of the moist 
thermometer did not depend on the form or size of its bulb 
and the manner in which it is moistened. I ascertained that in 
air little agitated, in the class room of physics of the Collége 
de France, which has a total capacity of about 600 cubic 
metres, a thermometer with a spherical bulb of sufficient size, 17 
millimetres in diameter, exhibited constantly a temperature of 
0°10 to 0°20 higher than that indicated by two thermometers 
with very long cylindrical bulbs, placed immediately by the side. 
In the open air, the difference was maintained in the same di- 
rection, but it became weaker. The spherical bulb of the ther- 
mometer which I employed in this experiment is much larger 
than the bulbs commonly are of the thermometers which are 
employed in the psychrometer; but I have chosen it so pur- 
posely, in order to increase the difference, if any existed. I 
think it may be thence concluded that the form of the bulb ex- 
ercises only a very feeble influence on the stationary tempera- 
