Hi Danny,

Sorry i'm so late. It's a great paper - you've done alot of work
and it shows! Anyway, i'd have a few comments and some suggestions
for changes in wording (which you may take or leave as you wish). 
Bruce found many of the same things, so i won't reproduce them here.

Abstract:

New far-infrared ... presented for the Key Insights -->
We present new far-infrared ... for the Key Insights

*** I like to avoid writing in the first person. 

sec. 1, par 1:

low-metallicity systems like the Small -->
low-metallicity systems such as the Small

*** Done

you could (should?) also mention the hypothesis of
Lisenfeld et al. (2002, A&A, 382, 860) which proposes an excess
of small grains with small emissivity index.

*** Thanks.  I added this reference where I discuss modified dust emissivities at the end of the paragraph.

sec. 1, par 2:

allows for an unprecedented look -->
allows an unprecedented look

*** Done

bridge an important wavelength gap -->
bridge an important spectral gap

*** Thanks

local universe, where their dust emission ... wavelengths. -->
local universe, in which dust emission peaks at
    or longward of $\sim$150\,\micron.
    
*** I used part of your suggestion.

sec. 1, par 3:

In this effort we present -->
Here we present

*** I avoid first person writing.

advantage of one of the key features to Herschel data: -->
advantage of one of the key features of Herschel data:

*** Done

Note: while i appreciate the preview and publicity you give
of other KINGFISH papers, it might be better to be more
conservative. Perhaps you could mention only Gonzalo's, Joannah's 
and Maud's papers since they are certainly going to be published
soon. Foyle's paper covers only a few KINGFISH galaxies since
they are combined with Chris Wilson's GTO sample; this could
be a problem. My paper is very very late, and Daniela just
told me yesterday that she's very behind too on the paper
front. Perhaps also Alison?

*** I removed the Foyle reference.  I'll see how the other co-authors react to this.  I know some referees detest preview references, but in a large collaboration I think they serve a purpose.

sec. 1, par 4:

perhaps you could add global:
we search for evidence for a sub-millimeter excess

*** Done

sec. 2, par 1:

Fifty-seven --> fifty-seven

*** Thanks

perhaps you could cite Rob:
span wide ranges in luminosity, optical/infrared ratio,
and morphology (Kennicutt et al. 2011).

*** I already did in the first sentence; that's probably enough here.

The luminosity range is a factor of 10^4 -->
The luminosity ranges over a factor of 10^4

*** Thanks

the range of optical/luminosity ratios covers a factor of 10^3 -->
the optical/luminosity ratios span a factor of 10^3

*** Yeah, I had the same thought, but I was trying to avoid a third use of "span" in the same short paragraph.

sec 3.1, par 1:

The description of the PACS imaging is incorrect. Please
see Rob's paper for the number of scans, etc.; these also
(like SPIRE) were divided into two groups, one faint and
one bright surface brightness as determined from the
level at the optical radius.

*** Thanks.  I tried to elucidate the observing strategy
    via inspection of the AORs, but I should have just
    asked the person who scripted the AORs!

sec 3.2, par 1:

SPIRE imaging data were taken in scan mode, -->
SPIRE imaging data were taken in Large-Map mode,

*** Thanks

sec 3.3, par 1:

"The sum of the sky area ... uncertainty in the sky level
to the overall error budget." 
sounds vague to me. It is not clear what you mean since
later you calculate sky uncertainty only taking into account
the single sky aperture size rather than their number (see
below for more thoughts/confusion on this).

*** See comments below.

sec 3.3, par 2:

yield negligibly different global fluxes. -->
yield consistent global fluxes.

*** Thanks

sec 3.4:

"The apertures are chosen to encompass all the light detected at 
every wavelength."

sounds a bit vague to me. How did you define "all the light"?

*** I added "by eye"

Moreover, when you go on to say "diffraction" (beam smearing?), 
it seems that "all the light" was not really all the light
(although this is probably quibbling). 

*** You are correct.  The wording has been improved (I hope!).

i agree with Bruce's
comment here, and wonder how you can accurately derive such
a correction with 3.6 micron. In fact, it sounds like the 
corrections could be just the difference between the stellar 
and dust distributions, rather than effects due to beam smearing?

*** I added Bruce's suggested proviso.

are formulated to counter this effect. -->
are formulated to mitigate this effect.

*** Thanks

The description of the errors confused me. In the previous
section, it sounded like you took a number of sky regions
(say Nreg) each with an aperture of Nsky pixels. Then 
sigma_sky is defined as "the standard deviation of the
surface brightness fluctuations." Does this mean *within*
each aperture or *among* the different apertures? The
former would be an estimate of small-scale fluctations
(how small depends on the aperture size) and the latter an
estimate of larger-scale ones, over the regions spanned
by the "set of sky apertures" defined in the previous section.

However, from Eq. 2, it sounds like you mean within an
aperture if i understand correctly that Nsky is the number
of pixels of _each_ sky aperture, and not the sum total
of Nreg*number of pixels in each sky aperture. 
Eq. 2 is saying (again, IF i understand correctly) that
you consider sigma_sky as the uncertainty _among_ the apertures.
But you can't really do this (i would humbly propose) if
you then include the 2nd term in the equation as the error
of the mean within the aperture, namely sigma_sky/sq(Nsky);
namely, you would need another sigma_sky _within_ the 
apertures so that you could add the two with the scaling of 
Eq. 2.

As you can see, i'm confused! You might want to introduce
Nreg, or add an explanation of exactly what you mean by
sigma_sky, Nsky, etc.

*** Thanks for pointing out this bit of sloppiness.  I meant
to convey that the uncertainty is based on the (larger-scale) 
fluctuations derived from all the pixels in all the sky 
apertures.  I have made an attempt to clarify this.  In the
first paragraph of Section 3.3 I have tweaked:

"The total sky area, derived from the sum of the areas from 
all sky apertures," 

and near the end of Section 3.4 I have clarified:

"based on the measured sky fluctuations and the areas 
covered by the galaxy and the sum of the sky apertures," 

and:

"and Npix and Nsky}$ are the number of pixels in the 
galaxy and (the sum of) the sky apertures,".

sec. 4.1, par 1:

Did you really correct the PACS and SPIRE fluxes for extinction? How
big are these corrections?

*** Good point.  That wording was probably just borrowed from my
SINGS SED data paper, where such corrections could be very 
important (e.g., UV and optical fluxes were in that paper).  I
took a peek and the largest correction was 0.4%, and all but five
of the corrections are less than 0.1%!  What do you suggest?  For 
now I added a footnote that the corrections are no larger than the
0.4% correction for IC342's 70um flux (that galaxy is at b=10 deg).

sec. 4.1, par 3:

The suite of ... Spitzer data allow for a direct -->
The suite of ... Spitzer data allows a direct

*** Thanks

sec. 4.1, fig. 2:

Please use larger points (Bruce already mentioned this :-)!
Also, Boselli et al. (2010, A&A Herschel Special Issue) plot
(your) models in the color-color plots to show a typical SED.
Would this be helpful (yours or Bruce's), perhaps in the 
right panel where there is more scatter??

*** Done

Additional figure?
You might want to consider plotting SPIRE color plots as
Boselli et al. (2010, A&A Herschel Special Issue); they
might help show the power of the KINGFISH sample as it
may span a larger range of flux ratios.

*** Yeah, I have tried other color-color diagrams.  Note
that half of Boselli's color-color diagrams involve non-
Herschel data; since this is a Herschel data paper, I'd 
like to focus on Herschel-based color-color diagrams.  
One that seems independent of the two I already show in 
Figure 2 is 100/250 vs 350/500.  It looks similar to the 
distribution in the lefthand panel of Figure 2.  

sec. 4.2:

In addition to NGC 4736, NGC 337, NGC 2798, NGC 4631, and 
NGC 4826 SCUBA data are in reasonable agreement with the
SED fits. Is there some reason these weren't mentioned?

*** I was strictly referring to only the subset of galaxies 
for which Draine et al. (2007) excluded problematic SCUBA 
data in their fits.  The implication was that the other 
galaxies with SCUBA data were ok.  I've attempted to make 
this clearer.

sec. 4.3, 2nd to last par:

dwarf galaxies like DDO 053 --> 
dwarf galaxies such as DDO 053

*** Done

sec. 4.3, last par:

Aniano et al. (2011), but in this presentation -->
Aniano et al. (2011); here

*** Done

sec. 4.5, par 2:

"dust opacity coefficient beta are allowed to vary."
Over what range? Do you restrict the beta values? 
(This is probably similar to Maud's comment...)
There is some evidence for larger beta values (in the MW,
and perhaps also in KINGFISH, at least that's what my
preliminary profile analysis suggested) which would 
lead to larger dust masses, thus potentially closer 
to the Draine models.

*** In my BB fits, beta was a free parameter (it wasn't 
restricted to any sort of range).  I've tweaked to
"allowed to freely vary", and I've updated this section
to reflect Maud's wishes to explain what happens when I
fix beta to 1.5 or 2.0 (there's not much change).

In Fig. 6, perhaps you could consider plotting the
mass ratios against some color, perhaps SPIRE 250/350
or MIPS 70/100? There could be some trend with temperature?

*** Good idea.  Turns out your intution is correct -- there's
a slight dependence on FIR color.  Nice.  I've replaced the
figure.

Last (and very very least), could you put me as L. K. Hunt,
rather than L. Hunt? (sorry, this is really trivial... :-)

*** Done

Anyway, this paper will be a fitting introduction to the
KINGFISH data in all their glory! Thanks for including me
in this first pass!!

*** Thank you!