(1) let's not say things like this "extends" the ISO work. Either we have something new to say or not, and I think we do. We are exploring parts of phase space they could not detect. Let's just focus on that, and give the ISO results a lot less emphasis. We can give credit to papers and such where relevant, but one of our major conclusions (abstract, summary, and in text) should not be that we are simply confirming and extending ISO work. In that vein, all work on AGN is really just an extension of the work of Seyfert and Osterbrock. --The "extends and solidifies ISO work" sentence has been removed from the abstract. (2) Although I may have been the person that suggested this, I think all the extranuc spectra get too much explicit attention in the paper, and we should instead focus on the nuclear spectra alone. That said, I still think we need to give the numbers (average or median) for the extranucs, and maybe put a special symbol or box in the diagrams that show where these things fall. However, if we leave them out individually, it leaves open the possibility of doing a separate paper on them. I think this is better for both reasons - we really don't need the info on every extranuc for the current paper, simply the locus in the diagrams, it helps focus the current paper on the nuclei more, and it leaves room for a paper that explores the line fluxes and ratios in the extranucs alone. You could lead this with Brent, for example. I would remove table 2, and just give the numbers for the locus or show it on the plots. --This issue was addressed/resolved during our weekly telecons. (3) why match our apertures to ISO-PHT ? Do we really need the ISO archival data ? Again, I think it takes the focus away from the SINGS data and our contribution to the topic. In some of the plots (e.g., Fig. 3) it is actually useful, because it seems like we explore more of the low OIV/NeII than most of the ISO data in the archive. However, in many of the other plots, the ISO archival data simply make things unreadable. The NeIII/NeII plot suggests that the samples overlap a huge amount. Why not just plot the SINGS data there ? For places where a direct comparison is useful, or you really want to make the point that the SINGS data explore regions of phase space not touched by ISO, show both samples and be careful about the aperture matches (like you already have). For the other plots, just show the SINGS data, describe the results, and simply say that similar trends were seen in Verma et al., or Sturm et al. or whatever it is. --I have improved the presentation of the plots to make them more readable/understandable. We iterated on this via email. (4) Don't credit Weedman with using the 6.2 micron PAH line. Just reference the earlier papers from the special issue. This was Henrik's idea, not Weedman's. Weedman was just following us. Also, the problem with the 6.2 PAH is that there are actually confusing absorption features underneath (water ice and hydrocarbons), however in the SINGS galaxies where the optical depths are low (compared to ULIRGs), we don't have to worry about these. So, the 6.2 line is preferred, because it is not on the edge of the silicate absorption....which is usually much more of a pain in the ass. --Understood. I have removed the attributions for using 6.2um PAH in diagnostics, and simply referred in a prior sentence to all prior efforts that use the 6.2um feature as a diagnostic. (5) explain why you are using the 6.2 EQW in table 2 and in the plots. It's because the EQW gives you the ratio of the PAH flux to the continuum strength and at 6.2 microns, the continuum is either the tail of the warm dust from the star-formation, or, includes hot dust from the AGN. So the EQW of the 6.2 line is very sensitive to the presence of an AGN (indirectly). --Done (6) I think if we are calculating our own optical line ratios based upon our data for the nuclei, we need to give a table of these lines. It's a great idea to make a classical optical BPT, but I think we actually should give a table of these values for the nuclei. --I reference Moustakas and Kennicutt (2006), in which those data are published. (7) You should reference the old Veilleux & Osterbrock (1987 ?) stuff for the dividing line between SB and AGN. It's not really from Kauffman or Kewley. They may have added shock models or larger samples, but Sylvain did this way back as part of his thesis. I referenced it in my 1989 paper. --Done (8) don't put in direct quotes (page 7). Try to just summarize this - again, it is not surprising that the low-ionization lines are strong in the AGN. --Done (9) Although it is interesting to show this, I'm not convinced that the SiII/NeII line flux ratio is really such a good diagnostic. If you didn't know what the sources were classified as optically, could you really tell from Fig.3 what you had ? In some cases, yes, but the overlap in starbursts and AGN is much larger in SiII/NeII than in OIV/NeII where they clearly separate. Just look at your y-axis scale. The separation is an order of magnitude in OIV but only a factor of two in SiII. My claim is that this is interesting, but not really a great diagnostic, unless SiII/NeII is larger than 2.0. --I have added "regions" to Figure 3, along with a table that quantifies how well this diagnostic separates types. (10) in section 4.4.2, you compare the Magellenic clouds to AGN. In both, the excitation is high and the gas is hot, but for different reasons - as you say in the next section. I guess I'm not sure why we are making such a big deal about this. Can't we just say that the magellenic sources have hot gas becuase the metallicity is low, while the AGN have hotter gas because of the ionizing source ? Granted the line ratios of AGN are not the same as the magellenic sources, but that's just because of the x- rays. --I'm unsure what to add/subtract to the text, based on these comments. Do you have a more concrete suggestion? (11) In general, as I said the other day, I think your error bars are really large - too large. In some of the plots the errors in the line flux ratios are factors of 2-3. Surely our errors are not that large ? --This issue has been addressed, per our email exchange and the updated versions of the plots.