Beware the nonuniqueness of Einstein rings

We explain how an approximation to the rings formed by the host galaxies in lensed QSOs can be inferred from the QSO data alone. A simple ring image can be made from any lens model by a simple piece of computer graphics : just plot a contour map of the arrival-time surface with closely spaced contours. We go on to explain that rings should be (1) sensitive to time-delay ratios between different pairs of images, but (2) very insensitive to H-o. We illustrate this for the well-known quadruple lenses 1115+080 and 1608+656.