[X] My Assistant

[David]

On a non-spherical asteroid or moon (which I suppose is most of them), is the direction of "down" simply toward the overall center of gravity of the entire body, so that, for instance, on a flattened spheroid, a wheeled rover near the equator would find itself on a downward slant, and could roll "downhill" from the equator all the way to one of the poles? Or on a dumbbell-shaped asteroid (assuming the center of gravity to be halfway between the two ends of the "dumbbell"), on proceeding from the ends of the asteroid to the center, it might find itself plummeting straight down as the slope of the asteroid momentarily coincided with a line drawn through the center of gravity? Or is it more complicated than that?