Dynamic Screening in Thermonuclear Reactions

It has recently been argued that there are no dynamic screening corrections to Salpeter's enhancement factor in thermonuclear reactions, in the weak-screening limit. The arguments used were: 1) The Gibbs probability distribution is factorable into two parts, one of which, $exp(-\beta \sum e_{i}e_{j}/r_{ij})$ ($\beta=1/k_{B}T$), is independent of velocity space; and 2) The enhancement factor is $w=1+\beta^{2}e^{2}Z_{1}Z_{2}< \phi^{2} >$ with ${< \phi^{2} >}_{k}={< E^{2} >}_{k}/k^{2}$ and ${< E^{2} >}_{k}/(8\pi)=(T/2)[1-\epsilon^{-1} (0,k)]$. We show that both of these arguments are incorrect.