Obviously it's impossible, but no one on a different forum understood why not in terms of forces. Also the solution doesn't need to be quantitative. 1. The problem statement, all variables and given/known data The one thing I'd change from the troll science picture below would be to change the balls full of air on a string to a flexible string with no balls of uniform thickness. Also assume that friction is negligible. http://files.sharenator.com/1288271730828_Lots_More_Physics_Troll-s943x507-102899-580.png [Broken] 2. Relevant equations Archimedes' principle force of buoyancy = (force of gravity)(density of fluid)(volume of object) 3. The attempt at a solution I realize that if the system rotates that its energy doesn't change, but at the same time looking at the forces, I don't understand why it won't rotate counter-clockwise. Please help me to understand this. 4. The solution explained Even when a fluid is at rest, its molecules are moving in all directions. But if it's at rest, just as many molecules are moving left as are moving right. Any solid in the fluid will feel a pressure on all its surfaces due to the impacts of fluid molecules. In the presence of gravity, the pressure in a fluid increases with depth. If the pressure were constant on all sides of an arbitrary volume of the fluid, then it would sink due to the force of gravity, and it wouldn't be at rest. Therefore for a fluid to be at rest, the increase of pressure with depth must perfectly balance the weight of the fluid. Any solid fully immersed in a fluid will be acted on by the force of buoyancy. This force is due to the increase of pressure with depth; the fluid molecules hit the bottom harder than they hit the top. If this force is greater than the force of gravity acting on the solid, it floats. However, if there is a fish tank full of water, with a cylinder that extends both above the surface and below the bottom of the fish tank (like in the problem in this OP), then the water doesn't produce a buoyancy force. This is because there is no bottom of the cylinder for the molecules of the fluid to impact; all the molecules do is hit the exposed sides of the cylinder (and compress it a bit). So Archimedes' principle does not generally apply.