If you spin up a flywheel by applying torque somewhere else on its axle, you are transferring energy along the axle from the point where the torque is being applied to the flywheel, which effectively adds to the mass of the flywheel. This means rather counter-intuitively that a tiny force has to be applied along the line of the axle to keep it from shifting slightly towards the point where the energy is being applied (by an amount such that the center of mass of the added energy and the original rest energy remains at rest). I analyzed this system many years ago in response to a posting by a Dr. Eue Jin Jeong, who had not spotted this possibility and instead thought that this was a means of shifting the center of mass of the system in violation of conservation laws, which would lead to what he calls "Dipole Gravity". (He ignored my analysis and carried on by writing books about it). I think I assumed a hollow tube for an axle and a structure built of narrow isosceles triangles carrying the forces along it, but I no longer have my notes from that time, and I don't have the patience to reconstruct the analysis at the moment. Does anyone know of a suitable reference for the way in which Special Relativity means that any sideways transfer of force (especially as part of a torque) also implies a sideways component of force corresponding to the sideways rate of transfer of energy?