WITHDRAWN FROM THE ACTION OF GRAVITY. 343 
§ 4. The principles established in the two preceding paragraphs having elimi- 
nated from the question of our meridian lines the complications which might 
have embarrassed it, we my proceed to deal more directly with the subject. 
In the experiments relative to the formation of the liquid cylinder between 
two solid rings (2d series, § 38,) when the upper ring has been raised so as to 
cause the mass of oil to lose ifs spherical form, but not sufficiently to cause it 
to assume the cylindrical form, we obtained a portion of a figure of equilibrium 
of revolution pertaining neither to the sphere nor to the cylinder; it was shown, 
moreover, that if, after having formed the cylinder, the separation of the rings 
increased, there would result another portion of a figure of equilibrium equally 
differing from the sphere and the cylinder, and which will of course be under- 
stood as being also one of revolution. In order to determine what the liquid 
figures, to which the portions in question pertain, would be in their completed 
state, let us first cite a new experiment. 
We take as a solid system a cylinder of iron of considerable length in pro- 
portion to the diameter, and supported by two feet made of wire of the same 
metal (Fig. 5;) let the length, for instance, be 14 centimetres, and the diameter 
2. 'This cylinder being carefully rubbed with oil and introduced into the vase, 
we bring into contact with it, midway its length, a sphere of oil of suitable 
volume. As soon as adhesion takes place, the liquid mass spreads upon the sur- 
face of the cylinder so as to envelop a part of its extent, loses the spherical 
form, and constitutes in the end a figure of revolution whose meridian line 
changes curvature in the direction of its two extremities, becoming at those two 
points a tangent to the generating cylinder.** The meridian section of the liquid 
figure and of the cylinder is represented at Fig. 6. 
§ 5. As we have shown theoretically, (2d series, §§ 6 dzs, 10, 18 and 20,) and 
have verified by many experiments, when the liquid mass adheres to a solid sys- 
tem which causes it to lose its spherical form, the only parts of that system on 
which the new figure of equilibrium depends are the very minute lines along 
which it is in contact with the superficial layer or stratum of the mass, so that, 
the system may in general be reduced to iron wires representing those lines. 
Now, in the figure which we are considering, the free surface of the liquid mass 
touches our solid cylinder along’ two circumferences perpendicular to the axis, 
and passing by the points a and 4; we may therefore readily conceive the en- 
tire cylinder replaced by two rings representing those circumferences, that is, 
with an exterior diameter equal to that of the cylinder, and placed vertically as 
regards one another, having between them the interval a4. It will be necessary, 
however, that the quantity of oil should be greater in order to supply the volume 
of that portion of the cylinder suppressed in the interior of the mass; it will 
require even a little oil in excess to furnish the substance of the two bases 
which rest upon the rings, bases whose surfaces, as we shall presently see, will 
be convex spherical caps. In order to avoid these last, which would needlessly 
complicate the figure, we may take disks instead of rings; then, in both cases, 
the figure will be entirely formed of oil, and it is represented in this state, in 
' *M. Beer (see note 1 of § 1) indicates the same experiment for verifying one of the results 
of his calculations; but I had employed it long before. 
