Response to referee B:
 
We want to thank referee B for her/his comments and the issues she/he raised.

  1) In my view the presentation of the results of the extraction of the
  twist-4 matrix element f_2^n and of the color forces should be
  extended. Given the fact that about half of the introduction deals
  with this topic and that the title promises information about color
  forces, it is not sufficient to tell the reader that the results are
  presented in Table II and then let him draw the conclusions himself.
  The results should be compared at least to those presented in Ref.
  [17]. The information should be given, why the new central value f_2^n
  = 0.073 \pm 0.040 (at Q^2 = 4.3 GeV^2) differs substantially from the
  previous value of 0.034 \pm 0.043. Is it mainly due to the change of
  d_2^n from the previous value d_2^n = 0.0079 \pm 0.0048 down to the
  new value d_2^n = -0.00035 \pm 0.00108 or due to new extracted values
  for the other quantities entering the analysis? The contents of the
  last two sentences of the paper should not only appear in the summary
  but should be mentioned before in an extended discussion of the
  results presented in Table II.

We completely agree with the first comment of referee B. We would have very much 
liked to expand on our discussion of f2n and present more details that went into 
the updated extraction; unfortunately, due to the space limitation associated with 
PRL, we were not able to include more. However, among other pieces of experimental 
results and methods that did not make it in this submission, we do plan to have an 
extensive discussion of our f2n extraction in the archival paper of this experiment 
which is currently under preparation. Furthermore, we would have liked for example 
to add the f2n value from A. V. Sidorov and C. Weiss (Ref. [65] in revised version), 
but they only listed a f2^(u-d) value, thus a comparison in this case must wait for 
the SANE experiment (http://www.jlab.org/exp_prog/proposals/07/PR-07-003.pdf) proton results.
 
Besides differences due to the a2n and d2n values, there are other quantities that lead to 
our f2n extraction being different relative to that found in Meziani et. al Ref. [17].  
Our extraction used more recent JLab E94-010 data from K. Slifer et. al (Ref. [63] in revised version) 
and included additional data from the JLab RSS experiment (Ref. [62] in revised version). 
It turns out that the f2n extraction is rather sensitive to the E94-010 data given its precision 
and is the predominant reason for the difference in the f2n extraction. Although the central f2n 
values differ by roughly a factor of 2, we note that our f2n extraction is still consistent with that performed in Ref. [17]. 

Additionally, the analysis presented in Ref. [17] mixed quantities at different scales. 
In Ref. [17], a2n and d2n were both at a scale of Q^2 = 5 GeV^2, while their f2n extraction 
was done at a scale of Q^2 = 1 GeV^2. Our f2n extraction was done at the same Q^2 scale as our d2n and a2n data.

  2) Only experts that are familiar with the JLAB experiments are able
  to relate the acronyms for the various experiments given in the
  legends of Fig. 1 and Fig. 2 to the references in the captions.
 
  In detail: In Fig. 1 five experiments appear in the legend, namely
  E142, E155, E01-012 (Resonance), E99-117, and E97-103, and in the
  caption references [24, 54, 55, 57, 58] are listed.
 
  Ref. [54] can be related to [E142] and Ref. [58] to [E01-012], but
  Ref. [57] refers to [E154] and not to [E155] and from the given
  information it is open, which of the two references [24] and [55]
  refers to E99-117 and E97-103. Similarly the relation between the
  three acronyms that appear in Fig. 2 and the references [23, 24, 58,
  59, 62] is unclear. Both [23] and [59] refer to [E155] and not to
  [E155x].
 
  The captions should be extended somewhat to clarify the relation
  between experiments and reference numbers such that the reader is not
  forced to search it himself in INSPIRE.

We have taken into account the comments of referee B concerning the clarity of the references. 
The figure captions now pair the experiment names used in the figure legends with their 
respective references. The reference to the E155x data (Phys. Lett. B 553:18-24 (2003).) in Fig. 2 
was taken from the E155x published papers website found here: http://www.slac.stanford.edu/exp/e155/e155extension/e155x.html.
 
  * page 2, right column, end of 2nd paragraph:
 
  Instead of ‘in the first moment of g_1, \Gamma_1.’ I propose to
  write ‘in \Gamma_1, the first moment of g_1.’

This has been changed according to the suggestion of referee B.
 
  * page 3, left column, 2nd paragraph, line 5:
 
  blank before ‘(E = ..’ missing (twice)

Fixed all instances.

  * page 3, left column, 3rd paragraph, line 3:
 
  Why ‘found’? Shouldn’t it be ‘used’ or ‘presented’?

The word 'found' has been changed to 'used'. 
 
  * page 4, right column, 1st paragraph, lines 6-7:
 
  This sentence needs rephrasing. A ‘geometrical overlap’ of a
  counter and a signal cannot form a trigger signal

This has been rephrased.
 
  * page 4, right column, 2nd last line:
 
  Remove period between ‘Eq. 8’ and the subsequent equation (9)

Fixed.
 
  * page 5, left column, 2nd paragraph, lines 4-6:
 
  Add blank before ‘(evaluated..’ (three times)
 
Fixed all instances.

  * page 3, left column, 2nd paragraph, last three lines:
 
  It is unclear, how the statement ‘that is about 3 standard
  deviations smaller than that reported by the SLAC E155x experiment’
  is related to the number d_2^n = -0.00035 \pm 0.00108 from this
  experiment and the number d_2^n = 0.0079 \pm 0.0048 from SLAC E155x.

This has been rephrased to specify that the 3 sigma difference is measured from our upper error bar to the lower error bar of E155x.
 
  * Table II: dimension of F_E and F_B and equations (6) and (7):
 
  The authors consistently use the dimension (GeV/c)^2 for Q^2. For
  consistency then equations (6) and (7) should contain the proper 
  factors of hbar and c that are needed to get the correct values and
  dimensions of F_E and F_B.

We have added a clarifacation about the units to the caption of Table II.
 
  * Ref. [17]:
 
  comma missing after ‘et al.’
 
Fixed.

  * Ref. [19]:
 
  remove ‘et al.’

Fixed.

Once again we would like to thank referee B for her/his comments and helping us put forth a more refined paper.