Hello. I have a formula for gravitational acceleration http://www.part.lt/img/eb4c24f748f6e9a2797ed354bfbe6051663.png And I have tested it by numerical integration. It looks it gives perfect results. For example 0.104 arc seconds per revolution for the Mercury precession. But I have no an idea how to calculate it analytically. Maybe this could be interesting to note: As you see dm can be positive (when mass increase) and negative (when mass decrease). Therefore additional force v*dm/dt acts against the changing of the velocity v. This is very similar to magnetic induction. Maybe even magnetic induction itself can be explained in the same way?