Abstract
A semiclassical discussion is given of a simple case of the Hanle effect when the population of the level of interest varies with time because of decays to it from higher levels excited by collisions. It is found that the width of the intensity-vs-applied-field curve is no longer simply related to the lifetime of this single level, but depends on the relative populations and lifetimes of all other participating levels. It is shown by examples that, unless these effects are taken into account, the lifetimes deduced from the measurements can be greatly in error.