It will be found, however, that the floats _f_ No. 1 and _f_ No. 3 do retain equal distances from the centre in all corresponding situations, for the resistance to their approach to the centre by buoyancy is the elasticity of the sponge at the extremity of the respective levers; and as this elasticity is the same in all situations, while this centrifugal force of the float _f_ No. 1 is equal to that of the float _f_ No. 3, at equal distances from the perpendicular, the floats _f_ No. 1 and _f_ No. 3 will, in all corresponding situations on either side of the perpendicular, be at equal distances from the centre. It is true, that the force by which these floats approach the centre of motion varies according to the obliquity of the spindles on which they work, it being greatest in the perpendicular position; but, as the obliquity of these spindles is the same at all equal distances from the perpendicular, and as the resistance of the ascent of the floats is equal in all cases, the center of buoyancy will evidently describe a similar curve on each side of the perpendicular; and consequently the equilibrium will be preserved, so as to leave a constant moving force at _x_, equal to the whole accumulation of water in the sponge. Nor will this equilibrium be disturbed by any change of position in the floats not immersed in the water, since, being duly connected with the sponge by the levers and plates, they will evidently arrange themselves at equal distances from the center, in all corresponding situations on either side. It may be said that the equilibrium of the band of sponge may be destroyed by its partial compression; and it must be admitted that the centre of gravity of the part compressed, according to the construction above described, does approach the center of motion nearer than the center of gravity of the part not compressed. The whole weight of the sponge is, however, so inconsiderable, that this difference would scarcely produce any sensible effect; and if it did, a very slight alteration in the construction, by which the sponge should be compressed as much outwards as inwards, would retain the center of gravity of the compressed part at the same distance from the center of motion as the center of gravity of the part not compressed. CHAPTER VII Liquid Air as a Means of Perpetual Motion A few years ago air was liquefied. This was accomplished by a very high compression accompanied by a very low temperature.