572 PROFESSOR DANIELL ON THE WATER-BAROMETER ERECTED 
We next examined the portion of oil and water which had been set by in a 
glass vessel for the purpose of watching any changes which it might undergo. 
This we found in a very different state. The stratum of oil upon the surface 
was rather more than an inch thick, and in this it differed from that in the 
boiler, which was not more than half an inch. The great body of it was per¬ 
fectly bright and pure, and did not seem, from its taste, to have undergone any 
change, or to have acquired any rancidity. At the point of contact with the 
water it appeared to have undergone change, and to be separated from it by a 
tough film of the same mucilaginous-looking substance which we had found 
in the boiler. Upon agitating the glass, this film could be bent upwards with¬ 
out breaking; and a kind of fold was made in it of so tenacious a quality as 
to be some time before it again accommodated itself to the level of the liquid. 
Upon examination with a lens, it appeared to contain minute air-bubbles. 
These air-bubbles may have originated from some decomposition of the oil or 
water; but they were by no means numerous, and it is not at all improbable 
that they were the remains of a thin stratum of air included between the oil 
and the water; as there would be no perfect contact between the two liquids 
near the surface of the water. We next placed the glass, with its contents, un¬ 
der the receiver of an air-pump, and upon exhaustion of the air these little 
bubbles expanded and seemed to lift the film in parts and to escape with some 
difficulty through the oil. No air-bubbles, however, were formed in the mass 
of the subjacent water; proving that the water had been, in this instance, pro¬ 
tected by the oil. Upon pushing the exhaustion to the utmost, a few insig¬ 
nificant bubbles were indeed extricated from a small flock of dust which had 
fallen to the bottom of the glass. 
A little of the water was then taken out of the boiler in a glass vessel, which 
still retained a thin stratum of oil upon its surface. Upon exposing this to the 
action of the pump, air-bubbles in abundance were extricated from the whole 
mass, and it swelled up so as nearly to overflow the vessel in which it was con¬ 
tained ; presenting a very marked contrast to the result of the previous expe¬ 
riment, and proving that the water in the boiler must have been strongly im¬ 
pregnated with gaseous matter. This examination took place on the 13th 
June, almost exactly two years from the completion of the water-barometer. 
Upon consideration of all the circumstances, we were of opinion that the 
