FREED FROM THE ACTION OF GRAVITY. 25 
completely filled ; a certain quantity of a mixture less charged 
with alcohol, and marking only 16° on the areometer of Beaumé, 
is then cautiously added. . This, from its excess of density, falls 
to the bottom of the vessel, where it spreads itself in a hori- 
zontal layer. The oil is then introduced, which, by reason of 
the small excess of alcohol contained in the upper mixture, de- 
scends through the latter till it rests upon the denser layer of 
the lower mixture, either in a single mass or in several partial 
masses (§ 4). This being so, we unite, if the case requires it, 
the isolated spheres into a single one; then we stir the liquor 
cautiously with a glass rod, so as to mix imperfectly the layer 
at the bottom with the higher layers, but without dividing the 
mass of oil, and the system is then left to rest. It will be seen 
that there must hence result in the alcoholic liquor a state of 
density increasing from the upper layers of less density to the 
lower of greater density than that of the oil; and that, in conse- 
quence, the mass of oil will necessarily remain in stable equili- 
brium with respect to the vertical direction, in a certain layer 
whose mean density is equal to its own. Now, in performing 
the operation with the necessary precautions, that is to say by 
stirring the liquid only a very little, then leaving it to rest to 
observe the effect which results, again stirring it and leaving it 
to rest, and so on; lastly adding, if necessary, a small portion of 
mixture at 16°, or of pure alcohol, according to circumstances, 
we easily succeed in causing the mass of oil to remain exactly 
at the desired height, and, as we have seen, without tendency to 
a change of position in the vertical direction*. In geometrical 
strictness, truly, this mass of oil cannot then be any longer per- 
fectly spherical: it must be flattened a little in the vertical di- 
rection ; but, if we have operated so that the increase of the den- 
sities is very feeble at the height at which the oil stands,—and 
we easily obtain that result by suitable trials,—the flattening in 
question is completely insensible to the eye, and the mass appears 
_ exactly spherical. 
For the experiments which we have to describe, the most con- 
venient diameter to give to the sphere of oil is about 6 centi- 
metres. We easily accomplish this by first forming a less sphere, 
* The different liquid layers thus superposed tend of themselves, it is true, 
to mix; but, as they are placed in the order of their densities, this spontane- 
ous mixture proceeds only with extreme slowness, and if requires a great many 
days for the liquor to become homogeneous. No inconvenience therefore re- 
sults from this for the experiments. 
