Fig. 1:  The caption ought to specify the relative sizes of the inner and
outer radii.  Since the outer radius is already there, only the inner one is
needed.

--Done

Fig. 4:  The caption should specify what the grayscale image is.

--I replaced the figure with something that does a better and more quantitative job of getting our point across.

Figs. 8-9:  I find the wavelength labels a little difficult to read.  It might
help to make the font bigger and to blank out the background around the label
so no stars show through.

--I fixed it up a bit.

sec. 3:  Have you made any color corrections or attempted to put the
calibrations on a common system (i.e., some instruments are calibrated against
a stellar spectrum and others against a nu*F_nu=const. spectrum).  It would be
worth saying here what you did or didn't do to the data along these lines.

--I added such a blurb in the second paragraph of Section 4.

p. 6:  Is it fair to describe IRAC as two "imagers?"  I would instead describe
it as a single imager, with two "fields-of-view."

--Fixed

p. 9:  Maybe I misunderstood, but I thought that the extended source catalog
extractions were only too faint for a subset of galaxies with low surface
brightness.  The statement in the paper suggests that this offset applies
generally, i.e., for all galaxies.  Is that the case?  If not, it would help
the reader to modify this paragraph to make that clear, e.g. by adding "below
a surface brightness of <X>" after "too faint."

--This has been re-worked a bit, and I hope it now makes our point clearer. 

p. 10:  Are you certain equation 3 is right?  Probably I'm just being dense,
but it seems like the term in the square root would just go to N_pix as N_sky
gets large.

--Since N_sky is approximately N_pix, I've replaced (N_pix + N_pix^2/N_sky) with just 2*N_pix.  To make things more clear, I've added to the text that N_sky ~ N_pix.

p. 13:  It might be worth noting that Engelbracht et al. (2008; their Table 9)
get a similar 0.23 factor for the ratio of 8um to 3.6um stellar flux, using
models over a range of metallicity.

--Thanks for reminding me!

==============================================================
Additional communications between Chad and me.

> Thanks.  I fit an unbroadened PSF to N253, and I think that's
> appropriate.  I was roughly able to match the wings of the PSF, but more
> importantly I simply scaled it such that the global 24um is
> approximately equal to the global 25um flux.
> 
> I've read and re-read what was done for M82 24um saturation in the SINGS
> DR5 documentation.  It seems the delivery did not involve any kind of
> PSF fit (broadened or otherwise).  Since I intend to publish approximate
> M82 24um fluxes in the LVL paper, should I go with the relevant
> paragraph in the SINGS DR5 doc, or is there another way we should word
> it?

You've read the documentation correctly, i.e., we did not deliver any
PSF-fitted data.  I tried to produce a merged model/data image for the paper,
but was never happy with the result and so we never published it.  I'll attach
what we came up with in case it's useful to you, but you can see it's not
quite right anywhere - i.e., the core doesn't look right, and the diffraction
spikes don't match.  It's possible we'll never get the spikes right, simply
because it's tough to compute the model PSF correctly that far out, but it
should be possible to do better in the core.