For pt.I see ibid., vol.18, p.2647 (1987). The authors present here an experimental study of the variation in the shape of Ta(100) 4f core-level lines with temperature. These spectra exhibit a broad structure in the higher-binding energy side of the ⁵/₂ levels which can be assigned to the satellite predicted in the preceding paper. Moreover, a theoretical simulation of the total spectrum including the surface, first-underlayer and bulk peaks compares very satisfactorily with the experimental data. In particular, the overall shape of the spectrum and the variation with temperature of the linewidths are well reproduced. The presence of the satellite due to electron-hole pair formation thus explains the apparent anomalies of the Doniach-Sunjic parameters of the Ta surface peak put forward in earlier experiments. Finally, they note that it explains also the apparent different broadenings of the two spin-orbit lines.