A general comment:  This may be a matter of taste, but I find the style of
referring to observations, results, etc. as being "in" some author as awkward.
I prefer to say that actions are performed "by" people, not "in" people.
E.g., "...the work by so-and-so (some year)..."

--Fixed

Tables 1 and 2:  It might be useful to the reader to indicate the aperture
over which these quantities are averaged.  (They are averages, right?)

--Done

Figures 2 - 6:  It might be useful to the reader to indicate what the error
bars represent.

--Done

Figures 3 - 6:  You might consider replacing the light green lines with a
darker color (even a darker shade of green would help, I think - e.g., I tend
to redefine green in my plots using R/G/B values of 0/0.75/0).  That color is
notoriously hard to see when printed in grayscale and especially when
projected.  (Take pity on the poor astronomer presenting your paper in journal
club and having to explain to the audience that, yes, there really are lines
on that plot...)

--Done

Figures 3 - 6:  You might want to tell the reader the source of the
classifications (seyfert, liner, etc.).

--I think I will stick with just explaining this in Section 5.1, and not the fig
  captions.

Figure 3:  The caption should explain the significance of the roman
numerals on the plot, as you do in Figure 5, by referring the reader to the
appropriate table and section of the text.

--Done

Figure 4:  This may be a matter of taste, but I think discussion belongs in
the main text, not in the figure captions.  The comment about the excitation
level of the Magellanic Clouds should be moved to the text, in my opinion.

--Fixed

subject headings:  You might also want to include "galaxies:  nuclei,"
"galaxies:  active," and "HII regions."

--Done

sec. 1:  I'm surprised you don't mention the series of "dwarf seyfert
nuclei" papers by Ho et al. in your introduction, which appear to me to be a
fairly classic set of references on optical AGN classification.  If you want
to add one of those papers, an appropriate one might be ApJS 112, 315.

--I've added this reference to the Introduction.  [I did already reference two other
  papers by these authors later in the manuscipt.]

sec. 3.1:  To me, "centered" implies a specific choice to position the
extraction aperture.  You only extracted spectra once, so you can only
"center" that aperture once.  I.e., you can't be centered both on the optical
coordinates and on the infrared peaks.  I found this paragraph very confusing
until I figured out you were trying to say that the optical coordinates often
coincide with the infrared peak, which is interesting and useful information
that ought to be stated clearly.  (E.g., "we find that the optical coordinates
generally coincide with the infrared emission peak" or some such.)

--Fixed

sec. 3.1:  In reference to the continuum fits, to what function does "first-
or second-order" refer?  A polynomial?  Also, how do you decide whether the
fit should be first or second order?  What effect does the choice have on your
results?

--I have added "polynomial" to the text.  I typically used 2nd order only for the
  6.2um PAH feature and the NeII 12.81um emission line, since these are located
  in regions where the continuum has curvature.

sec. 4.1:  Should you also mention [Fe II] 25.99um?

--I decided not to include it since i) we don't use it; ii) it's not 'prominent';
  and iii) it's not 'high ionization'.

sec. 4.2:  You may also want to cite Madden et al. 2006 (A&A 446, 877) and Wu
et al. (2006, ApJ in press) for relevant work on PAH emission in galaxies.

--Done.  I did refer to Madden a couple of sentences down, but I added the
  reference again.

sec. 4.2:  The equivalent width of the 6.2um feature will also be sensitive to
the relative proportions of starlight and dust emission from the galaxy.
(I.e., there may still be significant stellar emission at 6.2um).

--Correct.  Ideally we'd normalize by TIR, but the 23"x15" aperture makes this
  difficult.  We could also normalize by the LowRes 6-38um flux, but we don't have
  LongLow for extranucs nor archival targets (over half the sources in the figures).
  I have added to the text that stars contribute to the EW.

sec. 5.1:  I would like to echo Daniela's comments about the extinction.  In
the complex mix of gas and dust along the lines of sight probed by our
spectra, a single optical line ratio does not probe the entire dust column.
There's not enough information presented here to demonstrate that "no sources
appear to be heavily buried."

--Thanks.  I added the corroborating evidence from SINGS 24-Halpha comparisons to
  back up this claim.

sec. 5.2.1:  As discussed earlier, the equivalent width of the 6.2um feature
will also depend on the relative proportions of starlight and dust emission.
You might check whether the low EW targets also have a low IR-to-optical
ratio.

--Thanks.  They are not unusually low.

[MINOR]

Table 1:  "Markarian 33" really stands out as the only galaxy identifier
written out in full.  Why not use the commonly-used abbreviation "Mrk"
instead?

--No particular reason.  ;)

Table 4:  Why are there 3 regions "IV?"  Should the last 3 regions instead be
"IV," "V," and "VI?"

--Fixed

sec. 1:  "6300 A Halpha" -> "6300 A, Halpha"

--Fixed

sec. 3.2:  "decent" sounds a little casual in this context.  Something like
"reasonable" or "acceptable" might work better.

--Fixed

sec. 3.3:  "We do not"  -> "The SINGS project did not"

--Fixed