18 PLATEAU ON THE FIGURE OF A LIQUID MASS 
Well, experiment confirms all this in a complete manner. 
The mass of oil, whatever its volume, remains in fact suspended 
in the midst of the alcoholic liquid, and takes the form of a per- 
fect sphere. 
3. In order to obtain this singular result with facility, it is 
necessary to take certain precautions, which I will describe. 
The first concern the formation of the alcoholic mixture. 
The density of this mixture necessarily varies with the kind of 
oil which is used. For the olive oil which I employed, and for 
the purity of which I cannot vouch, the proper mixture marked 
22 degrees on the areometer of Beaumé. If therefore any one 
wishes to use olive oil, he may always consider the above value 
as a first approximation, and by successive attempts will bring 
the liquor at length to the exact point which it ought to reach. 
To accomplish this, a test tube is filled with the liquor, into 
which a little oil is afterwards poured by means of a long-necked 
funnel, which reaches about half way down the test tube. The 
oil, on reaching the liquor, forms a globule, to which a diameter 
of about two centimetres must be given, and which a little shake 
will detach from the mouth of the funnel if it does not detach 
itself. Then accordingly as this globule falls to the bottom of 
the liquor, or rises to its surface, we conclude that the quantity 
of alcohol of the mixture is too great or too small; we therefore 
add to this a little water or alcohol, taking care to stir it well, 
and recommence the experiment of the test tube. The same 
operations are repeated until the globule of oil remains sus- 
pended in the liquor, without appearing to have a tendency 
either to fall or rise. The mixture may then be considered as 
approaching very nearly the desired point. I say very nearly, 
for the globule of oil of the test tube being of small dimensions, 
has more difficulty in moving in the liquor than spheres of a 
large diameter, and it may seem to be in equilibrium of density 
with the surrounding liquid, whilst for a larger volume of oil 
this equilibrium does not exist. 
4, When the alcoholic mixture, which I presuppose to be con- 
tained in a large glass flask of the ordinary form, has attained 
this point of approximation, the next thing is to introduce the 
mass of oil. For this purpose the long-necked funnel which 
has been mentioned above must be again used, and this must 
reach to a certain depth in the liquor contained in the flask, 
Letting the funnel rest on the neck of the latter, we pour the 
