It is now believed by all scientific men that none of these things are possible. In the first place, it is well known that compressed air will do exactly the same work in regaining its former volume that was expended upon it to compress it, and this with absolute exactness. In compressing the gas with a piston the force exerted upon the rod to drive the piston must be sufficient not only to compress the gas but also to overcome the friction of the tight fitting piston, and further, if the pressure on the rod be removed, the expanding gas will deliver against the face of the piston exactly the force and energy required to drive the piston for the compression, but not all of this can be returned to any machinery driven by the piston-rod, for a part will be lost in the friction of the tight-fitting parts. Thus here, as elsewhere, there is an exact equivalent of energy a part of which is consumed in friction, and only a part available for returned motion. The same thing is true in compressing a bag, except that possibly the bending of the fabric is less resistance than the friction of the tight-fitting piston. Still, the bending of the fabric is some resistance, and consequently the bag so expanding cannot return all the energy required for its compression, the difference being the loss, however slight, in the bending of the fabric of which the bag is made. Again, let us admit that a dilated bag is easily immersed in water, and that if inflated with air there will be considerable tendency to rise, but how much energy is required for the inflation? It is manifest that if it is immersed the weight of the water and its consequent pressure will resist the attempted inflation, and must be overcome before the inflation is complete. The deeper the immersion the more the compression, and consequently the more work required for the inflation. If a bag having a contents of one cubic foot were immersed a mile in fresh water, and if it should be attempted to inflate it, the reader will perhaps be surprised to know that the inflation would have to be done against a pressure of substantially 2,400 pounds to the square inch. It is simple that the deeper the bag is immersed the more work it will do in rising to the surface, but it is equally plain that the deeper it is immersed the more energy is required for its inflation. In each case the work of inflating is exactly equal to the work returned in rising to the surface, and there is not one whit to spare for running machinery of any kind.