central river for automatic stabilization?

On Jul 26, 2014, at 8:54 AM, ronfox@... [ssi_list] ssi_list@... wrote: For those of us too lazy or ill equipped to do the the math, visualize a bucket of water situated on a spinning platter.
This is a poor model. Actually, not even close. In our case all of the water is constrained to the walls of the cylinder (locally two dimensional) and moves around to different parts of the cylinders walls as the cylinder wobbles.
As the platter spins the water will tend to flow up the rim of the bucket as it experiences greater acceleration. Thus I agree with Joe's assertion that the water in a river (or connected reservoirs) will tend to flow towards a point experiencing the greater acceleration, which is the point furthest from the COM. Thus the river would have a stabilizing effect.