1. p2, top paragraph, the second to last sentence starting with "Genzel.."
I would specifically say that PAHs are weak in AGNs.

--Done

2. p5. Section 4.1.  You might say that all the lines except SiII
are nebular lines from ionized gas, and that SiII comes from PDRs/XDRS (i.e.,
warm atomic gas) and ionized gas.

--Done

3. p6., section 4.4., 4th line.  You make it sound like the 54.89 eV photons
come from HII region cooling.  They come from the OB stars, of course.

--Done

3-4 lines lower.  I would add to "exceptionally hard radiation fields"
the phrase "such as arise in AGNs"

--Done

last line on page 6. In Figure 3, I do not see the "normal nuclei" points.
Also, you might add the symbols for the LMC, SMC and Galactic HII in the
lower panel.  I was looking for them in the upper panel with no luck.

--There are SINGS nuclei in Figure 3, and I've plotted all the SMC/LMC/MW
  HII points that exist.  The figures have been cleaned up some - please
  take another look.
  
4. In Figure 1, why the downward spikes near 10 um?  Some kind of noise?
  
--I have removed a couple of noise spikes near 10um.

5. p.7.  This is your big question.  On the SiII.  Note that the SiII
comes from a combination of HII regions and PDRs when you have OB star
excitation and from a combination of HII regions and PDR/XDRs when you have
AGNs.  The first point I would make is that I agree with your first
hypothesis--XDR emission in AGNs could dominate the HII regions,
and so the SiII/NeII ratio might be high.  This might be the
explanation.  I also note, however, that in the new Kaufman et al
paper, we should that the ratio of SiII(PDR)/SiII(HII) increases with
increasing density.  In fact, in low density HII regions, most of the
the SiII comes from the HII region, and not the surrounding PDR.  So
there may be a bit of a density effect here too.  AGNs may have their
emitting gas (HII as well as PDR/XDR) at higher densities (and in
higher radiation fields) than the normal HII/PDR regions of starbursts or
normal galaxies.  This would also increase the SiII/NeII ratio.

--I have added this to the discussion.

On the hypothesis that you give in italics as a question, I do not understand
why you have stronger PDRs in AGNs, unless it is this density argument
above.  It seems like we could look at PDR/XDR models and see where
most of the SiII is coming from.  There was a paper by a Dutch guy
(Meijerink or something like that, I think with Marco Spaans) in the
last 2-3 years who constructed PDR/XDR models and I think looked into
this question.  If the SiII comes mostly from the XDR, then your first
argument (the one you have in the paper above the italic part) is probably
the correct one.  If most of the SiII emission (when irradiated by
a power law source) comes from the PDR, then it might be the densit
argument.  I would note that for a power law source, you have more EUV
photons relative to FUV photons that for an O star (which usually has
a jump downwards as you go from FUV to EUV because of the opacity at
the Lyman limit).  Thus, you would expect PDRs to be stronger relative
to the HII region around O stars than around AGN.  So I think you
either need the XDR to make the AGN strong, or the density (i.e., 
higher densities around AGNs which makes the SiII stronger).

--I have added the gist of this to the discussion.

6. p.8, top paragraph.  I do not like the way we show Figure 4, but
make no speculation on what would cause the differences you note at
the end of that paragraph.

--Do you have a suggestion?

next paragraph.  You worry about using "X-ray dissociation region".
That is what Maloney and I called them, but the europeans and 
Dalgarno call them Xray dominated regions and it is ok with me
if you like that better.
  
--I have switched to "dominated"
     
7. Figure 6.  This may be an interesting thing to pursue. I note that
the ratio of SIII 18.7/SIII 33.48 gives you the electron density (see
Bob Rubins papers--Mike Kaufman can get you the reference--Draine has
shown the figure as well).  Thus, on this figure, you could plot straight
lines at 45 degree slopes that are contours of different electron densities.
You have almost two parallel, but offset lines, marked on the figure.  They
could simply be tracking different characteristic electron densities.

--Done