Observational constraints on spatial anisotropy of G from orbital motions

A phenomenological anisotropic variation \Delta G/G of the Newtonian gravitational coupling parameter G, if real, would affect the orbital dynamics of a two-body gravitationally bound system in a specific way. We analytically work out the long-term effects that such a putative modification of the usual Newtonian inverse-square law would induce on the trajectory of a test particle orbiting a central mass. Without making any a-priori simplifying assumptions concerning the orbital configuration of the test particle, it turns out that its osculating semi-major axis a, eccentricity e, pericenter \varpi and mean anomaly M undergo long-term temporal variations, while the inclination I and the node \Omega are left unaffected. Moreover, the radial and the transverse components of the position and the velocity vectors r and v of the test particle experience non-vanishing changes per orbit, contrary to the out-of-plane ones. Then, we compute our theoretical predictions for some of the major bodies of the solar system by orienting the gradient of G® towards the Galactic Center and keeping it fixed over the characteristic timescales involved. By comparing our calculation to the latest observational determinations for the same bodies, we infer \Delta G/G <= 10^-17 over about 1 au. Finally, we consider also the Supermassive Black Hole hosted by the Galactic Center in Sgr A^* and the main sequence star S2 orbiting it in about 16 yr, obtaining just \Delta G/G <= 10^-2 over 1 kau.