What Have We Learned About and From F_L(x,Q²) at HERA?

Recently the H1 collaboration has published a ``determination'' of the structure function F_L(x,Q^2) at low x. I address the question of how reliable this determination really is. I argue that it is in fact a consistency check of a given theoretical approach rather than a real determination of the value of F_L(x,Q^2), but potentially a very useful one. I compare the consistency of different approaches, and indeed find that a LO--in--\alpha_s calculation of structure functions is completely ruled out. I also find that the ``determined'' values of F_L(x,Q^2) are surprisingly stable under changes in theoretical approach but, when working consistently within a well-defined theoretical framework, the values of F_L(x,Q^2) implied are somewhat lower than previously quoted.