20 PLATEAU ON THE FIGURE OF A LIQUID MASS 
axis would be vertical, and enlarges in appearance the horizontal 
dimensions of the object. 
In order entirely to avoid this illusion, we must use a vessel 
of plane smooth sides, formed of plates of glass set in a metal 
frame (§ 8). We then have, in a complete manner, the curious 
spectacle of a considerable mass of liquid presenting the form of 
a perfect sphere, and imitating in some measure a planet sus- 
pended in space. 
Instead also of the above vessel, a glass balloon may be used, 
which is more simple and less expensive. In this case, indeed, _ 
the mass of oil only appears in its real figure when it occupies 
the centre of the balloon; but the apparent distortion is small, 
as long as the sphere does not remove considerably from this 
centre. A vessel of this kind is very convenient for most of the 
experiments which I shall describe in this part of the memoir ; 
but it would not serve for those which I shall have to make 
known subsequently. 
6. Now, having obtaimed, by means of the process above de- 
tailed, a fine sphere of oil well suspended, and presenting, I will 
suppose, a diameter of 6 to 7 centimetres, we shall observe the 
following circumstances, which it is important to notice before 
we proceed further. 
In the first place, the equilibrium, previously well established, 
is soon disturbed of itself: at the end of a few minutes we see 
the sphere quit its place, and rise with extreme slowness towards 
the upper part of the ambient liquid. If a little alcohol be then 
added to restore the equilibrium, on treating the mixture by the 
process of § 4, this equilibrium is again broken in the same 
manner at the end of a certain time. In fine, it is only by 
continuing for some days to maintain it by the successive addi- 
tion of small quantities of alcohol, that we come to obtain a per- 
manent equilibrium, which is then no further disturbed, except 
by an accidental cause, of which we shall speak in the following 
paragraph, If the temperature does not fall below 18° centigr., 
the above phaznomena are the only ones observed; but some- 
times, if the temperature remains below that limit, and always 
if it is below 15°, another effect is manifested, namely a dimi- 
nution in the transparency of the oil. 
These phenomena are owing to a gradual chemical action 
which takes place between the oil and the alcoholic mixture. 
The first of these would be very inconvenient in most of the ex- 
