I haven't done much more than quickly glance through the paper, but I do have a comment that I think could be a time consuming fix, so I thought I'd bring it up now.  I see that you've included the LVL logo as a figure, and while I like the idea, I have issues with the logo as it stands - the southern sky is simply a mirror image of the northern sky.  If we're going to include this in a refereed publication, I think the figure should either be the actual sky, or at least a different "representation" for north and south, rather than the same piece reproduced.  I've cc'd Janice on this e-mail, as perhaps she is in a
better position to redo the logo, or has suggestions on how/if to fix it.

--I've replaced it with the 'empty' version, per your and the team's recommendation.  

Authors:  Julianne's last name is misspelled.

--Thanks!

Intro, paragraph 1:  why are you singling out LITTLE THINGS as an example of a multi-wavelength data set?  I'd leave the e.g., as just SINGS, which is a well defined multi-wavelength study with a well defined representative sample.

--OK

page 5, typo near end of 1st paragraph:  ...Ge:Ga stimflashes do not occur...  (no --> not)

--Thanks

page 9, typo top middle of 1st paragraph:  ...each panel parameterizes the... (parametrizes --> parameterizes)

--I believe they are equivalent in English, but I will go with your preference.

page 10, bottom:  You comment that the 160-TIR ratios have smaller dispersions than 8-TIR ratios, but I was wondering how much of this is due to the fact that the low-metallicity (low luminosity) galaxies are not detected in the MIPS bands.  Is there a way to quantify this (perhaps adding a statement saying this is still true if one considers a restricted sample of galaxies with detections at all wavelength bands)?

--This is a fine point you raise.  Maybe I'm blinded by the established fact that 8um (PAH) is low for 160um, and ignoring that this may likewise hold at other wavelengths.  However, many of the low-Z systems are detected at 24um, and they follow the same trend exhibited by normal-Z galaxies.  Moreover, the normal-Z galaxies themselves show a much tighter distribution for 70/TIR and 160/TIR compared to 24/TIR and 8/TIR.  In short, while I think you made a great suggestion, I think the data show otherwise.

Figure 1: I do prefer the non-sky-filled version that Janice circulated this week.

--We've adopted it.

Figure 3:  Top panel is a repeat of Figure 2, but now shown with the x-axis reversed.  I am not convinced that this figure is helpful in characterizing which objects are non-detects; perhaps a plot of detection fraction vs morphological type would be more informative?

--I've switched the x-axis to match Figure 2, and I've also added absolute magnitude per Evan's suggestion.  My goal here is to show the sensibleness of our m_B~15.5 mag cut-off for the outer-tier of the survey--it's tough to have MIPS detections for these types of sources that appear in the inner-tier. 

Figure 4:  There is an 'a' in the lower left of the bottom left panel, which I don't think should be there.

--I've added all 26 parts of the figure now, and so it now makes sense for each part to have a letter.

Figure 8:  (doesn't exist...)  I'd like to see an additional figure that includes "most" of the galaxies in the sample.  Because you are using TIR or MIPS bands as one axis in most of the plots, there are a number of galaxies (mostly the faint dwarfs) that never appear in a figure.  I don't know if it is scientifically interesting, but perhaps you could add a figure with the IRAC colors [i.e., (1-2) vs (3-4)] so that even the MIPS non-detects appear in at least one plot.

--I've added a figure that shows the distribution of TIR, TIR/B and 70/160.

=========================================================

make continuum-subtracted ha for mosaic