Actually, the orbits precess so much that I suspect it might be inverse-linear gravity. (People trying to get inverse-square gravity right often do this by accident, because they forget or don't know that the unit vector in the direction of r, divided by r², is the same as the vector r divided by r³.)