In an attempt to assess the physical realism of the plage flux tube model derived by Bellot Rubio, Ruiz Cobo & Collados (cite{brc}, cite{bellot}) we extend its predictions to the Stokes V zero-crossing shifts of a number of Fe ii lines. For this analysis, accurate Fe ii central wavelengths are required. We have devised a procedure for bringing the available Fe ii laboratory wavelengths to the system of accurate Fe i wavelengths of Nave et al. (cite{nav}). It is shown that, relative to this system, the Fe ii system of Kurucz (cite{kur}) is shifted by 6.4 m Angstroms/ towards longer wavelengths. Some lines, however, are displaced by more than 10 m Angstroms/. Corrected central wavelengths have been used to extract the observed Stokes V zero-crossing shifts of 16 Fe ii lines. Comparison with the values resulting from the model of Bellot Rubio et al. (cite{brc}, cite{bellot}) suggests that the velocity gradients derived by these authors are esentially correct.