The attempt of Seeger and Frank1to reinterpret the annealing in recovery Stage III in gold in terms of the mobile single interstitial is discussed. Considering the experimental data of Stage III in irradiated samples and for the annealing stage observed in quenching samples around room temperature, it is concluded that the assignment of the recovery in these two stages to two different elementary defects is unjustified. It is further demonstrated that the interpretation of recovery Stage III by interstitial migration is in conflict with the direct observations of vacancies migrating in this stage and that the interpretation of the recovery measurements in radiation doped gold by the two interstitial model leads to inconsistencies. The attempt to interpret recovery Stage III in gold by interstitial migration must therefore be rejected and the activation energy of 0.71 eV measured for this stage must be attributed to the migration energy of single vacancies.