620 REGNAULT’S HYGROMETRICAL RESEARCHES. 
In order to obtain a current of air saturated with moisture 
at a fixed temperature, I employed at first two tubes in the 
form of.U filled with moistened sponge, which were immersed 
in a large vessel of water kept at a constant temperature. A 
thermometer was placed in this water, which was continually 
agitated. A second very delicate thermometer was cemented 
into the second tube containing the sponge, at the place where 
the aspirated gas issued; it indicated the temperature of this 
gas. . 
I found in this arrangement great inconvenience, which caused 
me to abandon it. When the air passes through the apparatus, 
there is always a perceptible difference between the thermometer 
placed in the current of the gas and that which is immersed in 
the water of the vessel, so that it becomes difficult to answer for 
the temperature of the gas and its state of saturation. In order 
to obtain constant numbers, it is necessary to draw up the satu- 
rated air into a large space where the air is apparently in repose. 
I have adopted the following arrangement :—A cylindrical tin 
tube of 25 litres capacity (Plate VIII. fig. 2), closed above, is 
placed on a large plate filled with water ; this cylindrical tube has 
three tubulatures. In the upper tubulature e is placed a very 
delicate thermometer, the bulb of which occupies nearly the 
centre of the vessel. In the tubulature f, the first tube contain- 
ing pumice-stone moistened with sulphuric acid is fixed, in such 
a manner that this tube draws up the air into the middle of the 
case. Lastly, by means of the third tubulature g, the cylinder 
is placed in communication with a balloon O filled with moist- 
ened sponge, which the air is obliged to pass through before 
reaching the cylinder. In order to be more certain of the state 
of saturation of the air, a cylinder of wire-work, enveloped inter- 
nally and externally with a piece of moistened linen dipping into 
the water which covers the bottom of the plate, is placed inside 
the tin cylindrical tube. A small aperture 9, in this case, allows of 
drawing up the air into the centre of the vessel, near the bulb 
of the thermometer. This apparatus is placed in a chamber the 
temperature of which varies little, and the experiment is not 
begun until some time after the apparatus is arranged. The ve- 
locity of the current was purposely varied, in order to ascertain 
whether this exercised a real influence upon the quantity of hu- 
midity found. Two experiments were made at the same tempe- 
rature; one with a current continued for forty-five minutes ; the 
