Evolving Newton's Constant, Extended Gravity Theories and SnIa Data Analysis

If Newton's constant G evolves on cosmological timescales as predicted by extended gravity theories then Type Ia supernovae (SnIa) can not be treated as standard candles. The magnitude-redshift datasets however can still be useful. They can be used to simultaneously fit for both H(z) and G(z) (so that local G(z) constraints are also satisfied) in the context of appropriate parametrizations. Here we demonstrate how can this analysis be done by applying it to the Gold SnIa dataset. We compare the derived effective equation of state parameter w(z) at best fit with the corresponding result obtained by neglecting the evolution G(z). We show that even though the results clearly differ from each other, in both cases the best fit w(z) crosses the phantom divide w=-1. We then attempt to reconstruct a scalar tensor theory that predicts the derived best fit forms of H(z) and G(z). Since the best fit G(z) fixes the scalar tensor potential evolution F(z), there is no ambiguity in the reconstruction and the potential U(z) can be derived uniquely. The particular reconstructed scalar tensor theory however, involves a change of sign of the kinetic term $\Phi'(z)^2$ as in the minimally coupled case.