Abstract
The Baldwin effect is a weak albeit significant anti-correlation between the equivalent width of high-ionization lines (typically C IV lambda 1549) and luminosity. However, the rest-frame equivalent width W(C IV lambda 1549) depends more strongly on Eddington ratio than on luminosity. We show in this preliminary contribution that the dependence on Eddington ratio may entirely account for the Baldwin effect (i.e., for the weak luminosity trend) in two idealized samples: a first one that is complete and volume-limited, and a second one that is flux-limited.
