1) In section 3. where the Milky Way foreground extinction is removed,
the extinction curve used is from Li & Draine (2001) [BTW, this
reference is missing from the bibliography].  This is a model extinction
curve.  Did you actually use the model extinction curve or the observed
curve they used?  If you used the observed curve, I think it would be
better to reference the paper that presented the observed curve.  I'm
betting that this is an R_V = 3.1 CCM curve [Cardelli, J. A., Clayton,
G. C., & Mathis, J. S. 1989, ApJ, 345, 245].

--I used the model.

2) In section 5.3, the discussion of the far-IR color includes a
reference to "clean lines-of-sight" for UV photons to escape.  "Clean"
seems a bit vague.  Maybe it would be better to say low tau (optical
depth) sightlines?

--Suggestion incorporated.

Also, I am wondering if the discussion in this section could be
tightened up.  I will think about this some more and maybe send you some
revised text if I can come up with some that makes sense.  :-)

===================================================================
1) In the introduction, you quote Gordon et al. 2000 as supporting the
use of the UV spectral slope for estimating the UV extinction.  This is
not what Gordon et al. 2000 found, but the opposite.  The main results
of Gordon et al. 2000 is that the ratio of IR/UV is a *great* measure of
the UV extinction except for systems dominated by very old stars.  Thus,
it would be better to shift the reference to this paper to the next
sentence talking about the IR/UV ratio.  You should also add this
reference (and Witt & Gordon 2000 for completeness) to the discussion of
why the UV spectral slope versus IR/UV is not a tight relation.

--I moved the reference to Gordon et al. 2000.

2) You mention that the MIPS fluxes and uncertainties have been tweaked,
but you don't say why.  The reason is that the MIPS calibration has been
updated and you can even cite the three relevant papers (Engelbracht et
al. in prep., Gordon et al. in prep., Stansberry et al. in prep.).  All
of these papers have drafts and will probably be submitted sometime in
the next month or so.  Just finishing all the details (which may change
the updates MIPS calibration factors - just a FYI).

--Done

3) In Tables 2&3, you list uncertainties for everything except for the
optical/NIR.  I noticed you had uncertainties in the ASCII tables you
pointed me to.  I think it would be great if you were to add the
uncertainties for the optical/NIR data in Table 2 - as you've already
done the work and it would help people using the data.

--The uncertainties in the optical and nir data are dominated by 
  systematics, so I prefer to simply state that percentage.

4) In section 5.3, Fig. 13 is introduced, but the introduction seems a
bit limited.  This section starts with a discussion of dwarf irregulars,
but the plot has all the data in it.  It seems like a sentence or two is
missing since Fig. 13 has all the data in it.

--Good point.  I was trying to bridge the discussion between the end of
  5.2 and the beginning of 5.3, but I can see that now I've sort of
  stunted the overall introduction for 5.3.  I've tweaked the text 
  to introduce all the data.

5) In the 3rd paragraph of section 5.3, the jump into describing the PSF
fitting is a bit abrupt.  Maybe something like "...dust from within thse
regions (Helou et al. 2004).  To test this idea, we have decomposed the
24um images into unresolved (point sources) and resolved emission.  The
point source photometry was done using..."  This might require a bit of
modification of the rest of the paragraph, but I think this would make a
better start to what we've done.

--Done

6) In Fig. 10, the caption lists 84% and 10% for the two eigenvectors,
but the text says 88% and 7%.  Which is right?

--88 & 7.  Thanks.