[Clascomment] OPT-IN:Di erential cross sections and spin density matrix elements for the reaction gamma p --> phi p

[Elton Smith]

The paper is nicely written and very informative.

I have a question for my own edification, which could also be used to clarify the procedure for the reader. Refer to the paragraph preceding  Eq. 5 on p. 6. This is the explanation for the event-based background separation approach leading to the
signal quality factor Q.  I don't think I understand the procedure. (Note: I also looked in Dey's thesis, which did not clarify the procedure either). My understanding is that one takes each event from the sample, one at a time. For each event, one generates 100 "closest neighbors" and forms the distribution in missing mass of MM(p). This distribution is fit to a signal and background function, which are used in the quality factor Q. My understanding is that the 100 closest neighbors are from phase space, and therefore all background. How do you end up with a non-zero signal function s(m)?

Minor clarification, p. 6 top right, first paragraph, parenthesis:
(with a mass-dependent width, explained above) -> (Eq. 9)

p. 10 bottom right, 
polarization axes -> quantization axes

p. 12 middle left column, it states
"In the case of the Schilling extraction using equation 33, the expressions are fit to a small subset of the data."  Do you want to add
a clarifying note of "...for the purpose of qualifying the PWA method."? Or is there some restriction on the use of larger samples with this method?


p. 19, caption Fig. 16. Next to last sentence about "....taken form the charged-mode topology"  It appears this (and there may be other references) may be a remnant of other text that discusses the neutral topologies? You may wish to omit them for clarity. 

p. 20, Fig 18. last sentence "...suppressed the older data" -> supersedes? 

This comment, together with the one phrase at the end of the first paragraph on p 2, leaves the reader hanging without much explanation about  the discrepancy between 2013 and 2000 CLAS data sets. From Fig. 18 it appears there is a factor of about 2 in normalization, which is hard to justify with "improved tools and better understanding of the detector." I assume this has been vetted extensively in analysis reviews, but it would be nice if better understanding of the discrepancy was given. 

p. 27, Fig 27. Enlarge the figure to fill the space.

p. 27, Section B and also abstract/conclusion.
The SCHC violation is summarized in Fig 27. It would be nice to give a number for the amount of violation. I assume maximum violation would correspond to rho00^0=1, so the plot indicates about a 10% violation. If such a interpretation is valid, it might be useful to present an average over a range of sqrt(s), which can be quoted as a conclusion. If not, one can always quote the average value of rho00^0 over the range. This could provide a quantitative number that can be quoted to strengthen the text in the abstract.