Brent's first round of comments were several handwritten suggestions 
that were mainly on written style of the paper.  These have been 
addressed.
-----------------------------------------------------------------------
Abstract:

I was confused by the $\gtrsim$90\% Seyfert/LINER nucleir or 
$\gtrsim$90\% H II regions... statement.  I think it's because of the 
three diagnostic diagram preposition.  It got me to think that there 
should be 3 regions.  Probably no one else would be confused by this, 
but it's something to think about.  Perhaps replace "In fact" to "Within 
three diagnostic diagrams"?

--It's been reworded

Introduction:

You mention the [O III]/Hb vs. [N II]/Ha diagnostic diagram here.  Do 
you want to move Fig 2. to Fig 1. and reference your new plot here?  
That way, the reader can see it at the very beginning. 

--I'd rather the first figure emphasize the infrared spectra, since
  that's the thrust of the paper.
  
...yet optical studies typically show a much smaller percentage of 
ULIRGS is powered by... should be replaced by either of the two choices
...yet optical studies typically show that a much smaller percentage of 
ULIRGS is powered by... OR
...yet optical studies typically show a much smaller percentage of 
ULIRGS powered by...

--Done

Section 2.1:

In the first sentence, would it be a good idea to reference the DR3 
documentation that is given online? 

--I thought about giving a link to the relevant Spitzer website, but
  my experience is that the SSC changes these links (some of my SSC
  bookmarks are already out of date).
  
Section 3.1:

For the Long Low spectra,  we're just showing them to be complete, 
right?  Because the statement in parentheses (Long Low data are not 
used...) makes it sound like it's not important to look at them.  I'd 
just get rid of the stuff in parentheses. 

--I thought of this too, but if I remove it then the reader may 
  wonder in the next sentence why I don't describe the LL mapping
  strategy.

Section 3.3:

I found this section confusing.  So, you are stating that EWs are not 
possible with Hi-res data because no background is subtracted.  So, we 
instead use the MIPS 24 micron fluxes as a makeshift (didn't want to say 
equivalent) EW.  Is that right?  The first two sentences of this section 
can be scrunched into one to say what I've just summarized above.  (if 
that's right, of course)
Oh, should this imaging section be placed right next to the Spitzer IRS 
data since it's being used there or placed after all spectroscopy 
sections.  It seems out of place after talking about the ISO spectra, 
but I understand your order as you have it.  Just something to think 
about. 

--This is one of those instances that I have tried to appease two 
  different reviewer's remarks.  I guess it's easier to keep things
  the way they are, rather than juggling three suggestions....

Section 4.1:

Here, I think that you mention X-ray dominated regions for the first 
time.  You might want to put a reference for this region here because 
not many general readers will know what it is. 

--Done

Section 4.2:

...PAHs for which the... could be ...PAHs where the...
..., there is no... could be ..., where there is no...
..., and is the not... could be ..., and where the feature is not...
Just trying to balance the conjunction between the three pieces. 

--I have taken your cue to clean up this sentence

Section 5.1:

...starburst galaxies systems.  should be ...starbursting galactic 
systems. OR ...starbursting galaxy systems.  OR ...starbursting galaxies.

--Fixed

...in the optical versus doing so in the IR,... should be ...in the 
optical versus in the IR,...
Just balancing both sides of the versus. 

--Done

Footnote 2: missing \AA after 6584.

--Done

Section 5.2:
(3rd paragraph)
I was wondering if [Si II] emission XDR ratioed to [Si II] emission from 
HII is greater than 1 for all densities.  Or is the similarity of the 
graphs due simply to a density effect?  Can we separate the two 
possibilities? 

--I found literature comparisons of XDR/PDR and PDR/HII, but not XDR/HII

(5th paragraph)
The Schaerer & Stasinska reference needs a year. 

--Fixed

Section 5.2.3:

Can't we get the metallicities from the SINGS archive site or 
Kennicutt's paper and determine the importance of metallicity?  At least 
for the nuclear points...  That is, is the change of [S III]/[Si II] 
normally so small between low-Z and high-Z systems? 

 a) John Moustakas (w/Kennicutt) is still working on the SINGS
    metallicities, using both his spectroscopy and values from the
    literature.  He has provided me with 38 nuclear and 51 'disk'
    (0.4R_25) metallicities.  36 of the 38, and 46 of the 51, span a
    factor of only 4 in Z.  This is still a work in progress.
 b) Unfortunately, we have IRS non-detections from most of our faint,
    low-Z galaxies.
 c) I don't specifically address metallicity in this paper, save for
    mentioning how SMC/LMC HII regions appear in various plots.  i.e., I
    don't use the SINGS metallicities in this paper.

Summary:
First paragh

I thought that you should change "large portion" to a fraction or 
percentage of the SINGS sample that you use.  Just a little more 
quantitative. 

--I want the sample to sound "large", and 2/3 nuclear and 1/4 extranuc
  doesn't give that impression

Second paragraph 

...[Ne II] 15... should be ...[Ne III] 15...

--Thanks

Figure 5.

I was confused.  Is the numbering scheme in Figure 3 the exact same as 
the one presented in Figure 5?  If so, great.  If not, then you may want 
to change one or the other to like letters or arabic numerals.