sec. 1:  I find the choice of opening paragraph a little odd.  There's nothing
wrong with the paragraph itself, it's just that it would make more sense to me
to start the introduction more broadly with a discussion of galaxy SEDs and
their applications (more topical for this paper) rather than a discussion of
dust properties and effects (which sounds like the intro to a dust modeling
paper).  You might instead open with a paragraph describing the general
characteristics of galaxy SEDs and the broad physics that drive them (e.g.,
UV-NIR is generally stellar emission, IR-mm is generally emission from dust
that has been heated by absorbing primarily UV stellar radiation, etc.).  You
could talk about various uses to which SEDs have been put, such as measuring
luminosities, constraining star formation histories, measuring dust masses,
modeling photometric redshifts, assessing nuclear activity, etc.

--I guess the reason I used this opening was because I wrote something along
  these lines for my previous SINGS SED paper.  With this intro I was trying to
  emphasize more the interplay between UV and IR emission, since the UV (and
  optical) data are what's substantially new to this work.  But I see your
  point--I am mulling over ideas for changing this.

sec. 2:  I'm not sure a single paragraph merits a section all its own.  This
discussion is given by Kennicutt et al. (2003), so you could direct the reader
to that paper and remove this section.

--I considered this, but others have cautioned me that they prefer to see such
  a description in the current paper, and to not have to look up another
  paper to understand the sample.

sec. 3.2:  You should specify the criteria used to identify foreground stars.
(I.e., what range in the ratios you list identifies a source as a star?)

--Some text has been added to this effect.

sec. 4.1:  You should describe, in general terms, the fitting procedure you
used to fit models to the data.  (E.g., was it least-squares minimization?)
You say you fit the dust model to the MIPS data.  What did you do where there
are no MIPS detections?

--I've added that a least squares fit was performed. I don't fit a dust curve
  if there are no MIPS detections, as shown in Figs 2&3 (i.e., m81DwA and
  HoIX).  The only exception is M82, for which I fit the model to the IRAS
  25, 60, & 100um fluxes (I've added that note to the text).

sec. 4.1:  You might also point out an additional benefit of your "standard"
stellar reference - it serves to highlight the relative importance of stars
and dust in each galaxy, particularly in the transition from stellar to dust
emission in the MIR.  For example, contrast a galaxy like NGC 628 (dusty) with
one like NGC 1404 (not so dusty).

--Done

sec. 4.2:  Daniela's 2005 paper might be a better reference for demonstrating
the correlation between 24um emission and star formation.

--Done

sec. 5:  The IR/UV ratio is not simply "a measure of the amount of extinction
at ultraviolet wavelengths."  As you point out in sec. 5.2, this ratio is also
sensitive to star formation history and metallicity.

--Text modified

sec. 5.1:  The claim that disk inclinations (even moderate ones) do not
significantly alter the optical thickness of galaxies doesn't make sense to
me, is not necessarily supported by Fig. 11 (see notes on sec. 5.3 and Fig. 11
below), and stands in contrast to other SINGS results (cf. Regan's NGC 7331
paper).  Perhaps you meant something more specific than the broad statement
made here?

--Your point is taken.  I've changed the conclusion to be less broad:
  "...moderate disk inclinations are not a dominant factor in determining the 
  infrared-to-ultraviolet ratio in SINGS galaxies."

sec. 5.3:  I'll buy the claim that Fig. 11 shows no obvious trend for dwarf
and irregular galaxies (see notes on Fig. 11 below), but I'm not sure that
result is necessarily interesting.  The inclination isn't always well defined
for those systems (i.e., they're called "irregular" for a reason).

--I've removed the irregulars from the plot and modified the text.  Thanks.

sec. 5.3:  You might also cite my 2004 paper on NGC 55 for a similar
decomposition of 24um emission into discrete and extended emission.  For that
very late-type galaxy, I think I found only 1/3 of the emission to be
discrete, but that number may not be directly comparable to your systems:  NGC
55 is edge-on and many discrete sources could easily be blended together.

sec. 6:  I'm not sure you've demonstrated that the specific star formation
rate is the "dominant regulator" of broadband spectral variations.  As you've
discussed, geometry can also play an important role.

--Good point.  I have added a new section on the specific star formation rate,
  and linked the text in Section 6 to this analysis.
  
table 2:
- Why haven't you made aperture corrections to the MIPS data?  They're not as
  large as for IRAC, but could still be significant. 
- The MIPS calibration uncertainties have come down significantly.  Why not
  use the new values?
- The "nodata" marker ("...") is used inconsistently - sometimes it means
  there really are no data, and sometimes it means there's no detection.  You
  should present upper limits where you have data but no detection.
- I prefer the term "calibration uncertainty" to "calibration error."  I.e.,
  if you know you've made an "error," why not fix it?
  
--These changes have now been incorporated.

fig. 1:  There are a number of galaxies which seem to have very poor fits,
especially in the MIR.  You might want to add a section which describes the
fits and reasons for the departures, either in general terms or perhaps more
specifically, with a subsection for each discrepant galaxy.  Obvious oddballs
include NGC 1377, Ho II, NGC 2915, IC 2574, NGC 4236, DDO 154, DDO 165, and
NGC 6822 (which has problems in the optical, too).

--Descriptive text has been added for the mid-infrared oddbals.

figures 11-14, 16:  The caption should state the source of the error bars.  I
assume they reflect the observational uncertainties in the tables, so you
could just say that.

--I added such notes to several figures.

figures 11-12:  I think figure captions should simply explain the contents of
the figure and leave any analysis to the text.  If you agree, the statements
regarding a trend with orientation (fig. 11) and the relative importance of UV
emission with respect to morphology (fig. 12) should be moved to the text.

--Done

fig. 11:  This looks like a scatter plot because galaxies of all types are
plotted here.  If you just look at the early-type spirals (S0-Sb), there are
some pretty clear trends with inclination.  In fact, it seems to me that the
more "organized" (i.e., earlier types are smoother and more aptly described as
a disk) the disk (i.e., the more meaning the inclination has), the stronger
the trend.

--Well, I could fit the individual trends, but by that point I will have
  very small number statistics in each bin.  Maybe we just don't have enough
  data to say anything here.


fig. 16:  The "infrared-to-ultraviolet ratio" is defined differently here than
in figures 11-14, despite being given the same name.  You should be careful to
distinguish the quantity used here by giving it a different name, perhaps
"infrared-to-FUV ratio."

--Done

MINOR:

sec. 4.1:  In a minor inconsistency, "alpha_SED" is listed as just "alpha" in
the plots.

--Fixed

sec. 5.3:  How can a "color" be "elevated?"  I think you mean to say the
70/160um ratio is elevated.

--Fixed

sec. 6:  "less then" -> "less than"

--Thanks

sec. 6:  "geomeotry" -> "geometry"

--Thanks

tables 1-2:  NGC 3034 doesn't appear in these tables, resulting in an orphaned
footnote in Table 2.

--This was intentional, but I've put it back in for you.

table 1:
- You might specify how PA is measured.  (I assume it's E from N).
- You should change the table note to a footnote on columns 3-6.
- A footnote on column 7 should direct the reader to the portion of the text
  which explains the resolved/unresolved ratio.
  
--Thanks

table 3:
- It would be nice to have uncertainties on the optical data.
- As in table 2, "..." is used inconsistently.
- Footnote "a" should appear in the table before "b."

--The uncertainty in the optical data is described in Section 3.2.  I 
  now include upper limits in Table 1, and would have done this for Table 2 but
  I do not have any upper limits to estimate---for optical data, "..." does in
  fact mean we don't have any data.

figures 1-9:  The light green and light blue colors are very hard to see in
grayscale or when projected on a screen.  You might change them to a darker
shade.

fig. 1:
- The caption should direct the reader to the portion of the text which
  describes the fitting procedures (which will be 4.1 if you follow my
  suggestion above).
- You should change "alpha" in the plots to "alpha_SED."
- How is the dust model fitting done when there are no MIPS detections?
- You should include MIPS upper limits on the plots.

--Done

figures 14-15:  The captions should direct the reader to the portion(s) of the
text defining and describing the "nuclear-to-total" and
"resolved-to-unresolved" measurements.

--Done

fig. 15:  Why are there no error bars on this figure?

--Error bars have been added.