Tom and I iterated quite a bit on IRAC extended source aperture
correction factors.  Below are a few more interactions.
================================================================
A bit hard to tell from the text if you are using 2MASS photometry,
but the SEDs clearly show the JHK measurements.  The NIR is important
since it nails the peak in the stellar distribution *and* it is
not subject to the wacky calibration issues that IRAC has
(nor is it extinction sensitive as the optical is).

So are you using 2MASS, and if so, presumably you are using
the LGA measurements?  You should have a reference in the paper to
this effect (Jarrett et al 2003).

What I'm really after is this -- which 2MASS measurements are you
using for the global measurements: isophotal or total mags?

--I'm using 'j_m_ext', 'h_m_ext', and 'j_m_ext'.  So total mags.

  In the text I point the reader to Dale et al. 2005 for details on ISO,
  IRAS, 2MASS, SCUBA, etc, which references Jarrett et al. 2003 and 
  talks about 2mass total mags computed by extrapolation out to 5 disk 
  scale lengths, etc.  But I'd be happy to put the Jarrett et al. 
  reference in this one, too.  Let's see, where could I put it?...it's 
  not obvious since I don't spend much text on 2mass this time around.  
  How about I add it to the Table 3 notes?
=============================================================
General comment about your TIR vs. UV correlation
analysis:  the global comparisons (i.e., totals, including
nukes, arms, bulges, disks) are a real mess, which again
confirms to me how complex even 'normal' galaxies can be
from a global standpoint.  It is truly a mixed bag by
throwing together all of the sings galaxies and their internal
constituents.  The casual reader will probably be horrified.
Nevertheless, you have pulled out a statistical difference
between active low mass galaxies and big old 'normal' galaxies.
I wonder what would happen to your plots if you segregated the analysis
to some real internal structures (nukes, arms, inter-arms, bulges,
outer disks).  This is of course a very difficult game with resolution
differences in serious play (particularly across Spitzer bands).  The nuke to
total analysis is a good start.  I'm just worried that the referee
will jump on this bandwagon.

--Yeah, it would be quite tough to decompose the 70 and 160um data
  into nuclear regions, arms, interarms, etc.  It would be possible
  for some of the nearest, largest galaxies, but that would diminish
  the sample size.  I've referenced a couple papers that have analyzed
  TIR/UV in terms of integrated vs local (end of Intro's 2nd par).

Let me jump to Figure 12 and your intrepretation.  For the early types,
notably the E's, it seems to me that the drop in the ratio is nothing
more than the simple lack of IR emission (i.e., the galaxies are mostly
dust free, except for some nuke emission due to the remnant of some long
lost galaxy eaten by the big elliptical).  These guys can have plenty of
UV radiation, but need not have any IR emission -- and hence, they are
totally uncoupled.  For the S0 types, the dust is quite wimpy, with
not much PAH popping up (nor thermal).  So you begin to see the break
down around Sa where the simple absense of dust compared to plenty of
UV is rendering the correlation defunct.  I thought your intrepretation was
a bit complicated, but then again I may be over-simplifying.  Just a thought
coming from another side of the fence.

--Thanks.  I mention low TIR/UV for Es is either due to UV excess or to IR
  paucity.  Maybe I should give more weight to the more likely latter scenario,
  but I guess I was influenced by reading a few GALEX papers on UV-conspicuous
  E oddballs.  OK, so I've now given the dust paucity more prominence in the text.

I don't really understand Figure 10 (or maybe the intrepretation).
You only see a real difference in the UV+optical, and hardly any significant
difference between the e-vectors in the NIR/MIR.  I would think you'd see
a difference in the thermal dust emission as well.  Is this because dwarf
galaxies and star-bursty little buggers have all this UV radiation (unquenched)
but little FIR absorption (due to the relative lack of dust)?  This is a similar
situation to what we see in the tidal tails of interacting galaxies
(see for example my recent paper on the Tadpole Galaxy).  Just a few sites
of massive star formation (the so-called super star clusters) can pump out
a lot of UV photons, but hardly any dust is hanging about (either destroyed or
blown away from the relatively wimpy environments).   These guys would stick
out like your Fig 10 examples.  John Cannon's work is another good example.
Still, I don't quite understand the IR side of the figure.

--I have rewritten this section in an attempt to be more illuminating.  [ You
  be the judge! ]

Ok, here are some suggested wording changes here and there.  Feel free
to take them as you see fit (i.e., they are, in most instances, just my
difference in word/style):

ABSTRACT --
eliminate "(e.g., optical morphology, disk inclination ... star formation history)."
Change the next sentence to
"Consistent with our current understanding of 'normal' galaxies,
the SINGS sample of galaxies in comparison to more active star-forming galaxies
exhibits a larger dispersion in the infrared-to-ultraviolet versus
ultraviolet spectral slope correlation."

--Done

The next sentence claims that early-types are to blame, and the next sentence
basically says the same thing. I would combine the sentences, maybe saying something
like "Early type galaxies, exhibiting low star formation rates and high
optical surface brightnesses, have the most discrepant IR to UV correlation ... etc"

--Done

Later in the abstract, eliminate "(smooth, clumpy or unresolved)"

-OK

Next sentence, change to "The thermal dust emission ..."

--Well, the 24um dust is perhaps semi-stochastically heated.

Last setence of abstract, change to
"For early-type galaxies with 24um nuclear emission, the dust temperature and
UV excitation are relatively high compared to disk galaxies."

--OK

Introduction--

change the first sentence to "Interstellar dust ..."

--OK

Second sentence, change to "Even after 80 years of study, the nature of dust emission 
features is still not well understood."

second paragraph, second sentence, change to
"Thus the combination of infrared and ultraviolet daa should provide a powerful
diagnostic of star-formation and selective extinction."

--OK

later in the same paragraph, change to
"One consistent result relevant to the work presented here ..."

--OK

later in the same paragraph, you have a typo, eliminate
"... , and (Bell et al. ...)"

--Thanks

Section 2, first paragraph, I would move the last sentence in the paragraph
to the first sentence; so it might read:

With median distance of ~10 Mpc and a max distance less than 30 Mpc, the 75 galaxies
in the Spitzer Nearby ..."

for the footnote that points to my IRAC calibration webpage, change to

1 From T. Jarrett's IRAC calibration page: http://spider.ipac.caltech.edu/staff/jarrett/irac/

--OK

Section 3.1, the second footnote is a bit ambiguous as to what it means.
do you mean that the method for reducing the data and extracting the photometry
is the same as for the GALEX Atlas of Nearby Galaxies?  So special processing
was required by Armando and company, right?

--'Yes' to the first question.  As for the second question, I thought it was 
  obvious from the text that the non-Atlas SINGS galaxies were treated in
  the same way as the Atlas galaxies.  Please help me re-word if you think
  it's necessary.

Section 4.2, third paragraph,second sentence, change "substantial" to "discernable"

--Why downplay?  The TIR/UV ratio in ensuing figures spans ~0.1 to >100.

Section 4.3, as I was saying above, I don't understand what is happening
to the long wavelength side of Fig 10.  You might add a sentence or two
of explanation to complete this section.

--I've tried.

Section 5.1, i was actually very happy to see that there is no correlation with
inclination, as i've been assuming this for years with my analysis
(that you can work with edgy galaxies and still learn something).  Nice job!

--Veronique had similar sentiments.

Section 5.2, as i was saying above, why not also mention that the correlation
is breaking down as the lack of dust (to emit or absorb) is coming into play
for these big old galaxies.

--OK, I've tried to make this angle the likely angle, giving it more prominence
  in the text.

also in this section, you have a typo in the forth sentence or so; change
"due excess" to "due"

--Thanks

Section 5.3, second paragraph, this section is hard to follow as I go back
and forth from the text to the figure;  I need a mental visual to help --- so
why not add some examples in here (i.e., name some galaxies that the general
discriptions that you are plodding through)?  Just mentioning a few galaxies
gives the reader a mental image (i.e., morphology) of what you are trying
to measure.

--OK.

Last sentence of this section is a bit of a weasel out.  Probably would read better
if you spelled out the "scenerio" that agreeing with your 24um morphology.

--I didn't like that sentence, and took it out; two sentences prior should suffice.

For the SSFR, I like this section and new paradigm for measuring SFR.
I seem to recall that Rob K. has a new body of work that post-dates the
1998 work ...?  Perhaps he has yet to publish the work.

--I hear it's coming soon.

The analysis and last sentence of the section 5.5 is really driving
home the point that early types are different from everything else because
of the lack of recent star formation, relatively high UV (massive systems),
and diminishing dust content.  Everything else in SINGS has new stars forming
(even if little wimpy dwarfs), molecular gas, and dust.

Which makes me think of stellar backbone mass.  You have not really addressed
that part of the sample.  You might find that the big massive galaxies also
exhibit the properties that you are seeing with the early types (yes, there
is overlap).  Of course, if a gas-rich massive galaxy is perturbed, it will
create a bomb of IR emission.  The sings sample is probably not big enough
to really treat the issue of mass and age (evolution).

Last thing, table 3, add a reference to the 2MASS Large Galaxy Atlas
(Jarrett et al 2003).  The K-band measurements are a crucial part of your
analysis, although this is not going to be appreciated by the typical reader.

--Done

Oh and one other thing.  figure 16, you should denote which points correspond
to the examples you show in figure 14;  that way the reader can see how the galaxy
color matches with the morphology (and visual image).

--Done