We are thankful to the reviewer for valuable comments and especially for the PRC 2013 reference which we didn't see before. Below we answer the highlighted issues. > - Introduction: the authors mention does previous experiments > significantly differ from statistical model > calculations. Statistical model calculations as such is somehow > meaningless. Which one ? With what input ? Was a sensitivity study > performed with respect to the various ingredients, i.e. gamma-ray > strength functions, nuclear level densities, nucleon optical > potential, etc.. ? To what extend is the conclusions from previous > experiments robust with respect to the various uncertainties in such > ingredients ? We added few words about the calculations in previous works, that are described in detail in the cited publications [1,2]. To save space (since the paper size is limited to 4 pages) we decided to include in Introduction only the final values that illustrate the amount of the observed difference between the expected and observed (g,n) and (g,p) yields. (g,p) reactions require additional effort for accurate description (e.g., [Tamae et. al, NP A690, 355 (2001); Shoda et al., NP A239, 397 (1975)]), so from our point of view it was enough to mention just "statistical models". At the same time further in this paragraph we removed the phrase that "statistical model calculations very often significantly underestimate photoproton reactions", since it would require additional explanation. > - The dipole strength function plays a crucial role in order to > understand the (g,n) and (g,p) reactions. However, the authors only > mention the reaction code (TALYS or CM) without any details of the > strength function considered. In particular, TALYS includes many > different models for the gamma-ray strength. More precision should > be given on the ingredients, but also on the sensitivity of the > predictions with respect to the different models. A short description of performed calculations was added following table 2, stating that different combinations of level density and SF models were tried using TALYS. Also a new comparison with the parameters taken from the PRC2013 paper was performed and included in the table. > - A priori, the relative (g,n)/(g,p) rate ratio may also depend on > the optical potential. A few word is needed ! Probably, the primary reason of the underestimated photoproton yield is the choice of optical potential. The global nucleon potentials used by the TALYS and CM models are different, but both do not take into account individual structure of this nucleus. It was mentioned at the end of Sect. 2 and in this version we also added a phrase to the Conclusion. Speaking about calculation of the corrected optical potential, it is beyond the scope of this short paper, so we limit ourselves to uniform normalization of the cross sections (after having included other possible sources of increased (g,p) reaction). > - The authors included the isospin splitting effect of the > IVGDR. More details are in order. We added a very short clause, that in this case microscopic CM calculation is used to provide the photoabsorption cross section for TALYS. There are more details of the model in the referenced work [9]. > - For most of the readers the resulting experimental bremsstrahlung > spectrum remains unclear. It would be important to illustrate the > energy region of interest. We discuss the bremsstrahlung spectrum in Sect. 3, and mention that it is wide in comparison with the Planck spectrum. There is no place to insert a figure showing the spectrum, but the electron energy and the raditor target that is used to produce it are described in the text. Also we inserted the Bp and Bn threshold energies at an appropriate place at the end of Sect. 2 > > - Sect. 3: the authors present nuclear network calculations and > their sensitivity to the (g,n) and (g,p) scaling factors. However, > as mentioned by the authors, the total photoabsorption rate is > actually rather well conserved and essentially the redistribution > between the (g,n) and (g,p) channels is not properly described, at > least at the energies of relevance in the present experiment. So, > the sensitivity calculations should take the constraints of a > constant total photoabsorption rate into account and the scaling > factors should be changed for both the (g,n) and (g,p) channels > accordingly. In this case, the 106Cd abundance would be > significantly less affected since the total destruction rate would > not be affected by the renormalisation. We agree that in the first version of the text it made an impression that simultaneous scaling of (g,n) and (g,p) resulted in 40% change of 106Cd abundance. We rephrased the Conclusion to mention that they are scaled separately. But it should be mentioned that in comparison with default REACLIB rates such a large difference does take place even when both are scaled, since there is no proportion relationship in this case. > - Sect.3: the sensitivity of the astrophysical rates around T9=2.5 > with respect to the various nuclear ingredients (gamma-ray strength, > level densities, optical potential) also need to be studied before > any renormalisation. > > - The argumentation that the normalisation factor applies for the > energies of relevance in the Planck spectrum is weak. The > enhancement of the (g,p) cross section might not be concentrated in > the high-energy part of the energy range, but still significantly > higher than the reaction threshold. More arguments should be > provided. In Sect. 3 an addition was made that we calculated the rates using different TALYS model parameters including the experimental ones from PRC2013 measurement. At p-process temperatures the resulting (g,n) rates differed very little, and the difference with the (g,p) rates calculated from experimental SF and level densities was <= 20% (even <= 10% at 2-3 GK). This is negligible in comparison with 400% corrections that follow from the measurement. Regarding the normalization step, we have deliberately chosen the initial cross sections so that the required scaling was as small as possible, so, as it is written in the text, the obtained values should present a conservative estimate. In any case we describe the shortcomings of the procedure and it is specifically stated that the obtained values are only an estimation and dedicated measurements are required. > - The 106Cd dipole strength between 2 and 8MeV has been studied by > Larsen et al. (PRC87, 014319, 2013) and can shed light on the (g,p) > threshold behaviour as well. Are the gamma-ray strength function > included in the statistical model calculation compatible with such > experimental data ? These results were used in the calculations as described above. > > - The authors never mention the alpha photo-channel which for a > Planck spectrum at T9=2.5 dominates over the neutron and proton > channels with its low Q=-1.626MeV threshold. If the (g,n) and (g,p) > rates are renormalised, why not the (g,a) !? The yield of the (g,a) reaction on 106Cd is not measured, since it produces stable 102Pd which is not accesible to the activation technique. > > In summary, I believe this paper is presenting new experimental > results but lack information for the readers, draw fast conclusions > on the impact on the p-process nucleosynthesis and does not properly > justify the rate renormalisation procedure proposed. Please provide > corrections along the above-mentioned comments before considering > publication. In addition, we have changed the title to remove 108Cd which is not described in this paper. Also a number of minor cosmetic changes were introduced.