Department of Physics2024-11-0920020217-732310.1142/S021773230200662X2-s2.0-0037034543http://dx.doi.org/10.1142/S021773230200662Xhttps://hdl.handle.net/20.500.14288/6706We argue that the geodesic hypothesis based on autoparallels of the Levi-Civita connection may need refinement in the scalar-tensor theories of gravity. Based on a reformulation of the Brans-Dicke theory in terms of a connection with torsion determined dynamically in terms of the gradient of the Brans-Dicke scalar field, we compute the perihelion shift in the orbit of Mercury on the alternative hypothesis that its worldline is an autoparallel of a connection with torsion. If the Brans-Dicke scalar field couples significantly to matter and test particles move on such worldlines, the current time keeping methods based on the conventional geodesic hypothesis may need refinement.AstronomyAstrophysicsPhysics, NuclearParticles and fieldsMathematicalJournal Article1793-6632175481600005Q22091