A Lagrangian model of vector-meson dominance similar to that due to Kroll, Lee, and Zumino is considered. It is shown that for a correctly diagonalized and renormalized Lagrangian the conventional diagrams of vector dominance cannot be simply interpreted as Feynman diagrams, and it is not possible to derive the usual current-field identity. However, it is still possible to derive $\ensuremath{\rho}$ universality, provided the photon and vector meson couple to the same conserved hadronic current, and also to derive a somewhat modified form of pole dominance consistent with a zero photon mass. The formalism presented here also provides a particularly simple technique for calculating the hadronic contributions to the magnetic moment of the muon.