Bulge microlensing optical depth from EROS 2 observations

We present two years of monitoring observations of the extremely variable quasar J1819+3845. We observe large yearly changes in the timescale of the variations (from ~1 hour to ~10 hours at 5 GHz). This annual effect can only be explained if the variations are caused by a propagation effect, and thus affected by the Earthʹs relative speed through the projected intensity pattern. To account for this effect, the scattering plasma must have a transverse velocity with respect to the local standard of rest. The velocity calculated from these observations is in good agreement with that obtained from a two telescope delay experiment (Dennett-Thorpe & de Bruyn 2001). We also show that either the source itself is elongated, or that the scattering plasma is anisotropic, with an axial ratio of >6:1. As the source is extended on scales relevant to the scattering phenomenon, it seems plausible that the anisotropy is due to the source itself, but this remains to be investigated. From the scintillation characteristics we find that the scattering material is a very strong, thin scatterer within ~ten parsecs. We determine a source size at 5 GHz of 100 to 900 microarcsecs, and associated brightness temperatures of 1010 to 1012 K.