We compare the predictions of a range of existing models based on the Vector Meson Dominance hypothesis with data on e + e − → π + π − and e + e − → µ + µ − 1 Permanent address. 2 Present address.cross-sections and the phase and near-threshold behavior of the timelike pion form factor, with the aim of determining which (if any) of these models is capable of providing an accurate representation of the full range of experimental data. We find that, of the models considered, only that proposed by Bando et al. is able to consistently account for all information, provided one allows its parameter a to vary from the usual value of 2 to 2.4. Our fit with this model gives a point-like coupling γπ + π − of magnitude −e/6, while the common formulation of VMD excludes such a term. The resulting values for the ρ mass and π + π − and e + e − partial widths as well as the branching ratio for the decay ω → π + π − obtained within the context of this model are consistent with previous results.