CWR QCD-10-035: comments' answers (for originals see http://cdsweb.cern.ch/record/1371874)

In black are the initial comments, in blue are the answers, in green when it needs discussion between the authors, in red when we have not adopted the suggestion (with explanation as to why), in pink when we have at least partially adopted the suggestion (with explanation as to why ),in brown when we need clarification of the comment.

Draft including the corrections

https://twiki.cern.ch/twiki/pub/Main/CwrDiphoton/QCD-10-035-2011-10-06.pdf

https://twiki.cern.ch/twiki/pub/Main/CwrDiphoton/QCD-10-035-2011-10-14.pdf

Legend

BLUE: Comment taken into account or answer to question

GREEN: Still to be discussed among us or to be fixed

RED: Not done, and we explain why.

PINK: Partial agreement with the comment. Changed done for the agreed part.

BROWN: The comment needs further clarification to be answered.

Proposed new plot style

Different comments suggested to change the differential cross section plots: luminosity error, legend, theoretical uncertainties appearing only in the theory/data ratio plots. Although the existing plots are public and already have been shown in conferences, in collaboration with the ARC we have developed a new plot style for the paper, taking into account the different comments. This new style can be viewed in the latest new draft intended for the FR (draft v1 in CADI which is the same as draft 2 on CDS).

Comments from the Alabama group, 25 Aug 2011

Dear QCD-10-035 Authors and ARC members,

Congratulations on producing a nice analysis and paper. Attached are the comments from the Institutional review by the University of Alabama group. I realise that these are after the deadline, so I guess you are not obliged to incorporate them, but I hope you will consider them anyway, as they are intended in the spirit of improving the paper.

Regards, Conor Henderson for the University of Alabama group.

QCD-10-035 University of Alabama Institution Comments

L7-12: I think you can say this a bit more concisely, and perhaps with a bit more 'punch', by something like: "A comprehensive understanding of prompt diphoton production is also important as it represents a major background to certain searches for rare or exotic processes, such as the production of a light Higgs boson, extra-dimension gravitons, and some supersymmetric states."

Done

L14: CMS -> 'Compact Muon Solenoid (CMS)', since this is the first time you mention it in the paper L15: LHC -> 'Large Hadron Collider (LHC)', for the same reason

Done

L31: 'For massless particles, Et is equal to pt ...' - I think this is pretty obvious to any HEP journal reader, and probably can be omitted.

Sentence changed to "Its transverse momentum is denoted $p_T$, $p_T = p \sin\theta$."

L47-50: A small comment: here you mention the full acceptance, but later you will show results from both the full acceptance, and only the central region. Somewhere in the paper you should explain this, otherwise it is a little odd when the reader first encounters it without any explanation (for example in Table 1).

Done

L57-58: Do you want to give the pt thresholds for the trigger objects? That may also help the reader understand better the choice of 23 & 20 as the offline cuts

Thresholds has been varied during the 2010 data taking period. Maximum value of the HLT Pt threshold has been added.

L102-103: You may want to mention if there is or is not significant pt or eta dependence to the photon id efficiencies

Dependencies are taken into account in the term (Nreco/Ngen)(Xi) from simulation whereas the efficiencies ratio is a single global correction factor data/MC.

L116: 'on a events' -> 'on events'

Done

L117: 'a direction' -> 'an axis' maybe?

Done

L147-162: I think this description could be a little clearer. Do you mean that you measure the iso for prompt photons from the random-cone method, then cross-check it with W and Z electrons? If so, then I suggest making it all one paragraph, and starting something like: "Contributions to the value of the ECAL isolation variable for prompt photons come from pile- up and underlying event activity. Since these contributions are independent of azimuthal angle, f(I) is estimated using the random-cone technique defined in Section 2, from events with at least one photon candidate. The probability density function obtained in this way is cross-checked using low-radiating electrons (defined in Section 2) from Z and W boson decays." and then continue with the Z and W descriptions as you have them.

Done

L148: 'independent on' -> 'independent of'

Done

L148: 'these contributions ... is' -> 'these contributions ... are'

Unchanged: Verb reports to f(I) not to "these contributions"

L165-166: suggest a slight rewording to the more concise: " In simulated events, this difference is found to be smaller than ..."

Unchanged: here we talk about another difference (MC photons Vs MC random cones whereas the first is data random-cones Vs data W and Z), so it should probably be specified

L214: 'uncertainty form' -> 'uncertainty from'

Done

L214: 'is of' -> 'is'

Done

L215: 'associated to' -> 'associated with'

Done

L221: 'are of' -> 'are'

Done

L222: 'and of' -> 'and'

Done

Table 1: As I mentioned before, I think it would be helpful to state somewhere that you will present the results for two different eta ranges, before encountering it here.

Done. See Alabama L47-50.

L240: I think the Theoretical Predictions section might benefit from a short introduction. Perhaps something like: "Theoretically, sources of prompt diphoton production are : ..." and then go through the contributions in order . Then you can say how each contribution is calculated.

Done

L241: Note that you also need to explain the fragmentation photon process here

Done

L246-248: Suggest a slight rewording to: "Although technically higher-order processes, the box contributions are quantitatively comparable to those from qqbar annihilation in the mass range of interest (including the region pertinent to the H->gg search), due to the significant gluon luminosity in this range at the LHC."

Done

L251: Suggest slight rewording to: "To simulate the effect of the experimental cuts, an additional isolation requirement at the parton level ...." And then I would not have a new paragraph between this line and line 254, since its really a continuation of the same concept.

Done

L253: It looks like the parton level iso cut at 5 GeV is looser than the experimental cut (which was 2 GeV)? Is that correct? If so, it needs some comment...

The cuts used at generator level to compute the efficiencies match the cut used in the theory prediction: 5 GeV.

L259: 'associated to' -> 'associated with'

Done

L260: 'and from' -> 'and'

Done

L273: I would suggest to re-list which variables you study here

It seems unnecessarily heavy.

Figs 2-9: Did you consider to show other theory comparisons here? For example, the Tevatron papers on this topic compared to Pythia and RESBOS as well as DIPHOX. Admittedly the Pythia comparison, being LO, would not be expected to be particularly accurate, but it still might be interesting to show, as it can indicate the kinematic regimes where higher-order contributions are most important. I also recall a very nice theory talk from Siegert in the QCD Photons meeting in May 2010, where he talked about Sherpa and photon production, and showed at the time good agreement with the D0 results. Did you consider to compare with Sherpa? Even you choose not to show other theory comparisons (I leave that to authors+ARC to decide) it might be interesting to at least show how the different sub-processes contribute to the overall 'theory' line ?

For ResBos and SHERPA, we were indeed interested in trying to compare with it, and some work was done in that direction. However, some technical problems were encountered which could not be resolved within the time remaining before the approval. SHERPA will be used for the gamma + jets cross-section paper which is in preparation on the 2011 data, and we would indeed like to use it for the next diphoton paper.

We think that showing the different sub-process contributions would be more relevant for a theoretical paper. Showing the different sub-processes seperately is not typically done in experimental prompt photon papers.

L 286-287: the phrase '2->3 at one loop and 2->4 at tree level' sounds very jargony. I hope you could find a better way to express this ...

Done: removed.

L298: just a small suggestion that the phrase 'using respectively random-cone and impinging-track methods' is maybe a bit too detailed for the conclusion, and that it would be slightly better to just omit it, keeping the rest of the sentence.

Done

Comments from the Baylor group, 15 Aug 2011

Dear authors and Roberto,

We have been assigned for the institutional review of this paper; however, due to miscommunication within the group, we didn't manage to send in our comments. We are very sorry for being late, but we thought we should still send our comments to this review HN. We are aware that these comments are late, so we understand even if these are not considered.

Best Regards, Hongxuan Liu and Kenichi Hatakeyama (for the Baylor group)

++++++++++

Congratulations for bringing this very interesting measurement to this stage. We find that the paper carries very interesting results and is generally well written. We have just a few analysis and text comments which you may find helpful to further improve the paper.

] L4: issuing from --> coming from ?

Done

] L10: comprehension --> understanding ?

Done. Part rephrased.

] L90-93: We find the choice of the word "acceptance" including the photon reco and iso efficiencies a little awkward. The unfolding correction may be more appropriate.

Done, rephrased simply to "correction factor". Actually, it is not the unfolding we are talking about here, unfolding is described in section 4.

] L95: Definition of N_reco_sim: Why both generated and reconstructed values of the observable X within some range? It seems to us that the more appropriate choice is to require only the reconstructed values to be within a range, as in the actual measurement you can use only the reconstructed values. Or are you making a matrix? In that case, do you want to have another index j to allow the off-diagonal elements, even if at the end the off-diagonal elements are negligible.

The migration between intervals X_i of the observable is corrected separately: see last paragraph of the Section 4. Both generated and reconstructed values of the observable X are required to be within the range to avoid double correction.

] L116: "a events" --> an event or events

Done

] L121: Therefore, quantities such as the number of impinging tracks or energy deposits in the isolation area are assumed to be the same as for isolated photons. How is this assumption justified? Line 119 says "The cone is also required not to include photon and electron candidates, nor jets.", but this must depend on how you define photon, electron candidates, and jets. If a study to confirm this assumption is done, can it be mentioned?

By definition tracks and energy deposit, apart from brem, around an isolated photon comes from underlying events. The shape of the isolation area having been chosen to exclude tracks and energy deposits from brem., tracks and energy deposits will be the same for random cone and isolated photons. The effect of definition of electrons, photons and jet in the veto on the presence of such objects in the random cone is of second order and such effects are taken into account by the ratio eps_data/eps_sim.

] L148: I don't think the f(I) is defined before it appears (though we know its meaning).

Done

] L163: What is the rationale of taking the difference between the distributions extracted from random cones and those obtained from Z and W events for systematics?

The idea is to extract the relevant f(I) distributions from data. For this we use random cones in photons events. To validate this, we use control samples of objects close to the photons we are interested in: low radiating electrons from Z and W. In the W and Z samples, we compare f(I) obtained in Random Cones and f(I) obtained from electrons, and we take the maximum observed difference as an uncertainty. This has been rephrased to be clearer.

] L180: How is f(I) weighted? We understand that it's in a way to reproduce the distributions of eta, n_vtx, and E_T of the diphoton sample, but can the procedure be explained more in detail?

We used distributions ratio for the reweighting. We think it is a technicality which is not relevant for the paper.

] L190: We are probably missing a point, but can't you set Ntot to be the same as N?

Here it is what people usually call an "extended" maximum likelihood function, taking into account the poissonian nature of the total number of events. A "basic" likelihood would be: L= product_i=1^i=N { sum_t_inT [(N_t/Ntot) f_t (I1, I2)] } with Ntot= sum_t_inT (N_t) if you multiply by a poissonian factor corresponding to the probability to observe N evts when you expect N_tot: N_tot exp(-Ntot) /N! you get the extended likelihood expression

Also, we suppose that this likelihood fit is performed for each distribution and each bin separately. Is that correct? We might have missed it, but we didn't find a place that this is explicitly mentioned.

The likelihood fit is done in each observable bin separately. This is indicated lines 197-199 on draft 0.

] L200, What is the estimated amount of biases? Can you give some numbers on the bias to give some idea to readers? We actually didn't know that maximum likelihood method has this kind of drawback for a small number of events

Done. Added: "It is less than 10% of the statistical error in 80% of the bins and never exceeds half the statistical error."

] L204: What is this "(2)"?

Done. 1 is the section number. It is specified now.

] L203-207: Can you define the response matrix earlier? Actually, we are a little confused. You say "It has also been unfolded for the detector resolution by inverting the response matrix"; however, it reads that the acceptance correction defined earlier in Eq.(1) already accounts for it.

The "Acceptance factor" includes only the effect of the finite resolution on the acceptance. Sentence L92-93 on draft 0 was modified to clarify this point. Definition of the response matrix has been added in the text.

] L228: the numbers 0.01, 0.03, 0.05. Are they percentage or absolute numbers?

No it doesn't mean %, those are actual variations of the pdf first bin. In percentage it would be ~1% for signal as the first bin is ~0.9, and ~10% for bkg as the first bin is ~0.3.

] L240-248: we think some information described here are good for introduction, for example the fact that the main signal processes qqbar->gammagamma and gg->gammagamma, and they are comparable in size in the mass range of interest (with reference to DIPHOX and Gamma2MC papers), so to give some insight to non-experts.

We have preferred to not enter in too many theoritecal details in the introduction and we have just indicated that the photons are produced in hard scattering of quarks and gluons.

] L275-288: We do not find any discussion on cos-theta* plots. Do the enhancement in data at high |cos-theta*| benefit from some interpretation?

We added a discussion of the cos(theta*) distribution. We've discussed of the discrepancy with the theorists E. Pilon and J.P. Guillet (Diphox authors). They explained us that it can be due to an underestimate of the fragmentation term, but they didn't have a definite answer for the explanation. We have just mentionned the discrepency in the paper without advanturing us in providing a theoritical explantion.

We have added a sentence in the caption Figure 10, to mention that the last bin is only populated up to 0.95 due to the pseudo-rapidity accceptance, as we think this precision precision was missing.

] Table2-5: We find the "acceptance" is misleading if we follow the content, as the acceptance was previously used when the correction factor is introduced. The "-" sign of some of the table contents are connected with previous digits. Overall, the tables may benefit from some beautification work.

Done. Removed "acceptance" and made tables more beautiful.

++++++++++

Comments from the FNAL group (Chetluru), 12 Aug 2011

Dear Authors,

Congratulations on first results of the two photon cross-section from CMS. Attached are the comments from the Fermilab Institutional Review. Please contact

Richard Cavanaugh : Richard.Cavanaugh@cernNOSPAMPLEASE.ch

Vasundhara Chetluru: vasurtang@fnalNOSPAMPLEASE.gov

for clarifications.

Vasu

GeneralComments1.docx.txt

Fermilab Institutional Review of CMS Paper QCD-10-035

"Measurement of the Production Cross Section of Pairs of Isolated Photons in pp collisions at sqrt(s)=7 TeV?

General Comments:

1) Comments should be written at the level of introductory detector physics, such as explaining why photon candidates are rejected by the presence of a track in the cluster cone.

We don't understand to which "comments" it refers to.

2) The order of the sections should be revised. For example, consider putting the Theoretical Predictions section after the Results section.

The theorical prediction section comes before the results section in order to introduce the prediction calculation before the results plots are shown, as those include the theoretical prediction.

3) The text is well written but sometimes a sketch or drawing would be useful to explain a set of cuts. For example:

a) page 2, line 80-83, sketching it was very helpful, even if it?s not straightforward to represent it well

b)page 4, lines 139-143, same thing

c)page 7: under theoretical conditions Feynman diagrams for the various channels would be nicer than many words of description

We have tried to keep the paper consice and we have limited the number of figures.

4) Tables 2-5 are good to see and confidence-building.

Specific Comments:

1) Include the integrated cross section in the abstract.

It was actually included in the early draft of the PAS, but it was decided to remove it during the review process, considering the result not general enough to be quoted in the abstract. Indeed the cross-section is integrated in a limited acceptance.

2) Include an explanation in the text of how the 5,977 events break down between signal and background, and among regions.

Break down bein tween regions done. Break down between signal and background was not added in this introduction as it is a result of the analysis.

3) The photon-photon angle in the Collins-Soper reference frame is introduced in line 45. There is no apparent connection to the formula giving the transverse momentum of the photon pair (line 43), nor any other explanation why this parameter is relevant/significant.

Improved. The rapidity is defined on line 30 of V0. Added a reference to Bern et al. paper dealing with observables presented interests for Higgs searches.

4) What are the L1/HLT thresholds for the diphoton trigger? These should be described in the paper.

Done. Thresholds has been varied during the 2010 data taking period. Values of trigger threshold have been added.

For your information, the trigger stream used are:

HLT_Photon17Isol_SC17HE_L1R --> L1_SingleEG8

HLT_DoublePhoton17_L1R --> L1_DoubleEG5

HLT_DoublePhoton15_L1R --> L1_DoubleEG5

We have not distinguished L1/HLT thresholds in the paper as the L1 thresholds are superseeded by HLT ones.

5) Line 126: "acceptance correction factor ~ 81%". Is this the combined expression in Eq.(1)? e.g. does it include the epsilon_data/epsilon_MC ratio? Is the discrepancy mostly due to resolution effects or different reconstruction efficiencies between MC and data? Not clear to the reader.

It includes both factors. It is mainly due to the efficiency of the photon identification and isolation.

6) The section 3, lines 132-143 need to be reorganized. The introduction of the isolation variable "\cal{I}" has to be done in a separate paragraph, followed immediately by its definition (it is now pushed down to lines 138 and later) and explained clearly that it is the main discriminating variable. As someone not familiar with this variable, I would like to understand why \cal{I} is different for signal and for background, and to see the expected distributions (maybe with a S-over-B ratio plot), before I am introduced to Fig.1 with the final fit in line 146, or I am presented with the experimental techniques (in the second half of the page).

We are missing the point about reorganization. The whole paragraph is actually about the isolation variable \cal{I} and its starts with an introduction explaining why it is discriminant.

We have added the plots of the \cal{I} distributions fro signal and background (Fig. 1 of new draft)

7) Several programs are quoted for the theory predictions, and different PDFs are used, but in the figures the theory is shown as a line with no uncertainty. Please address this inconsistency.

The uncertainty on the theoretical prediction appears on the TH-EXP comparison plots.

8) The statement that agreement between data and theory is satisfactory is not entirely supported by the Fig 6a and 7a with factors > 2 off. Unless the theory uncertainty is a bigger factor, but this is not shown.

Done, we removed the sentences about agreement except an "overall agreement" quoted for the mgg spectrum.

9) Photons from pi0 and eta decay are not non-prompt. Better terminology is "direct" for what you call "prompt" and indirect or "from hadron decay" for all others.

We have now defined what is called "signal photon" and "background photon" at the beginning of the paper and we refer to these terms throughout the paper

10) In the paper, integrated cross section is discussed before the differential cross section, and logically the other way round would be better, and then to discuss the integrated theory x-sn after the data result.

The theorical prediction section comes before the results section in order to introduce the prediction calculation before the results plots are shown, as those include the theoretical prediction.

Following to your comment, we have moved the value of the integrated cross-section predicted by the calculation from the "Theoretical predictions" to the "Results" section when comparing measurements and predictions. We have added a sentence at beginning of the "Theoritcal predictions" section introducing this section in the context of comparison of measurements and theory.

To compare theory and data spectra (presented as absolute differential cross-section) it is good to know how their integrals compare to each other. For this reason we prefer to show first the integrated cross-section figures.

11) Different theory cross sections should be distinguished, they are not all 52.7 +- x +- y. ???What does this comment refer to????

Only one computation has be made. The two different codes, DIPHOX and Gamma2MC, were used to calculate different subprocesses and the result corresponds to the sum of the contributions obtained with each of the two codes.

12) It does not make sense to write a syst uncertainty as (eg) +0.0035 - 0.0034, as if we knew anything that well. Better to say +/- 0.0035 (e.g.) for all cases where the difference is not important.

What does it refer to? Is it for the tables? In this case: we chose to keep 2 columns for syst. for consistency, as some of them are not symmetric.

13) Similarly Table 1 gives different syst uncertainties for the diffl and integral cross sections, but they are the same, as near as makes no difference, so we should just give one (even if it means rounding 6.3% up to 7%).

We have dropped the "integrated" column. \sim signs were removed in response to another comment and an explanation that they are only approximate values was added in the table caption

14) On all figs, the lumi error is indicated as a green band but is not on the plots. Better to take it off and just give the 4% in words.

Done.

15) Does the discrepancy between data and theory for the relative azimuthal angle in diphotons persist in the 2011 data? (It seems strange to be publishing a 36/pb result at this late date.)

The analysis was not done on 2011 data. The discrepancy observed on 2010 data is significant and is therefore expected to persist.

16) Indicate the average number of primary vertices per event, in the 2010 dataset.

Added to the text in parenthesis: "(2.4 on average)"

17) The Results section (7) could use some additional description and/or reorganization.

It has been modified considering others comments

18) Lines 163-166: Do we expect (e.g. from MC) that the Z/W and random-cone techniques would yield the same ("perfect") result? If yes, please say so. If not, why is the difference taken as a systematic uncertainty?

This has be rephrased to be clearer. The idea is to extract the relevant f(I) distributions from data. For this we use random cones in photons events. To validate this, we use control samples of objects close to the photons we are interested in: low radiating electrons from Z and W. In the W and Z samples, we compare f(I) obtained in Random Cones and f(I) obtained from electrons, and we take the maximum observed difference as an uncertainty.

19) Lines 172-176: The paper describes the experimental method for the determination of background \cal{I}: use an event sample with one-and-only-one impinging track. Makes sense. Then it proceeds by saying we are going to recalculate things with a sample containing TWO impinging tracks, and the difference in the result will be the systematic uncertainty. Why? Does the MC suggest that the two methods would give the same/similar results?

Rephrased to be clearer. Here we compared the two impinging tracks sample, with one track excluded from I computation, with the one impinging track sample, where the track is this time included in the computation of I. These distributions being similar, it validates the track removal method and the observed remaining differences are taken into account as a syst.

20) Use the same notations as in previous Photon papers, and be sure to be consistent with eta notation/ limits though out the paper. Sn is a new variable compared to the script used #sigma_{i#eta,i#eta}

Done

21) The technique used to obtain the signal fraction is quite new. Perhaps a better closure using MC study is ????will be interesting.

It was not done because we did not have enough background MC. Nevertheless, the fit algorithm was intensively tested with toy MC studies, and the several steps for PDF extraction were also tested separately on several control samples.

22) Explain e>300GeV choice for the isolation cone.

300MeV was chosen to minimise the dependency on energy deposited by MIPs such that the distribution for the bkg can be obtained from data. Specified in the text.

Line-By-Line Comments:

Abstract:

center-of-mass -> centre-of-mass (UK spelling convention)

Done

4] and elsewhere: 'Direct" photons is better than "prompt" photons .. as most photons from hadron decays are really prompt.

We have now defined what is called "signal photon" and "background photon" at the beginning of the paper and we refer to these terms throughout the paper. Both terms prompt and direct are used in the literature. Previous CMS papers on inclusive single photon differential cross-section measurements used the term 'prompt'.

4] diphotons --> photons (diphotons are defined later)

Done

5] testbed --> test

Done

6] approaches, as

Done

6] Prompt photon pairs

Done

9] are a major source of background, which in most cases -> are a major source of background, although in most cases

Kept unchanged as 'yet' follows

10] Yet a

Done

12] enhance the sensitivity

Done

15-18] can be rephrased

Was rephrased.

18] data set

Done

21] delete "immersed"

Done

27] A minor point, but "with respect to the beam axis" does not fully specify how eta and theta are referenced.

Done

37] along --> in

Done

41-52] A bulleted list may be easier to follow.

Done

46] "In addition" --> "In addition, "

Done

50-52] Specify that the isolation requirement refers to the parton level. (Or drop sentence)

Done

53-54] A GEANT4-based [7] description of the CMS detector has been used throughout the analysis whenever simulation studies are mentioned, if not otherwise specified> -> Unless otherwise specified, all simulation results are based on a GEANT4 [7] modelling of the detector.

Done

60-64, 69-76] Text is in repetition (and somewhat inconsistent) with similar description in lines 47-52.

Text is now consistent.

We did not remove the repetition as the first paragraph (47-52) is the description of the acceptance defining what is measured, and the second one (60-64 etc) is the description of offline selection. We think it is clearer to keep it that way.

62/63] 1.56 < |eta| < 2.50

We don't understand this comment

63-64] " the candiodate are required to be separated by R=0.45 to avoid overlap between their isolation region" and what if it is bigger than 0.45 ? ????above "iodate"????

Done

69-70] It would be good to give the logic associated with the cut, e.g "the sum of the transverse energy deposits in HCAL be less than 2 GeV"

Could you please be more specific about that? What do you think is missing?

69-83] The word "cone" is mentioned several times. It would be good to clarify somewhere that the cone is defined wrt/around the photon direction.

Done

69-72] The isolation criterion is different from that given on lines 50-52 ? How does the 5 GeV isolation requirement at the parton level (also given on 50-52) relate to the data criteria given on lines 69-72?

The cuts used at generator level to compute the efficiencies match the cut used in the theory prediction: 5 GeV.

79] define impact parameter

We added that it is the impact parameter wrt the beam spot, but we did not define it further, as it appears it is not done in other CMS papers.

84] The electron contamination is reduced -> The electron contamination is further reduced {Electrons have already been screened at some level by rejecting clusters with impinging tracks.}

Done

93] The N values in equation 1 are based on the SAME integrated luminosity (=?).

Both N refer to the same MC sample. This was rephrased to be clearer.

98] "described(?" in Section 1. "defined at generator level in Section 1" doesn't sound right.

Done

101] "epsilon_sim" --> "epsilon_data"? -->Simulated here refers to the reconstruction level.

Done

102] photon candidate --> photon PAIR candidate? OR to observe a photon - ->to observe a diphoton

Done

107] remove "working point 80% defined in"

Done

116] delete "a" ? "on a events" -> "on events"

Done

126-127] define two non-disjoint regions, |eta|<1.44, |eta|<1.44 || 1.56<|eta|<2.5 for the diphoton candidates. Since these are referred to extensively in the paper, this would be a good point to call this out and also explain the motivation for the division. Or perhaps this should be at line 50?.."|\eta|<2.5" is used as a shorthand in plot labels for the second region.

Done

129] delete "events with"

Done

129-131] "Main background ...coming from hadron decays". Is it events with multiple photons? There is a brief explanation about two collinear photons appearing as a single photon, but I do not understand how this is relevant to the signature. Where is the second photon?

The second photon is either a prompt photon (gamma+jet event) or another photon candidate from a jet (multi-jet event). Text rephrased.

130] delete "two"

Done

130] For neutral hadrons, the pair of two collinear photons can be misidentified as a single one.?

Yes.

The paragraph is concerned with background events and a hadron decaying into a coalesced photon does not contribute to the background?

Comment not understood.

135] ionizing -> ionising

Done

149-150] Not sure if there is a sufficient motivation to take the random cones from events with single photos. Have you compared the UE distribution for these different kinds of events in MC?

PDFs obtained from random cone on a single photon candidate sample and on the double photon candidate sample with data were compared. They are identical within the statistical uncertainties.

177-178] There is a mention of the distribution, but there is no distributions. Reword?

The distributions have been added.

179-181], It would be good to see a plot of the typical distribution f(I) for the diphotons signal and for a "typical" background and mention the pileup values encountered. (that might be doable for 7 pileup but will be a lot more difficult with 14 and certainly with 30!)

Same as previous comment

180] $\eta$, $n_{vtx}$[,] and $E_T$ serial comma rule

Done

187] How are the photons distinguished in equation 2 - does "1" refer to the photon with the higher E_T?

Done

195] For the 5,977 events, how do these divide between barrel-barrel, barrel-endcap, and endcap-endcap?

Done. Rephrased: "depending on whether both photons are in the barrel (2191 events), one is in the barrel and the other in the endcaps (2527 events), or both are in the endcaps (1259 events).

195] Rewrite to avoid starting the sentence with "5977"

Done

200-202] Need to quantify correction in the fit due to "small number of events".

Done.

204] identification efficiencies (2) -> identification efficiencies not sure what (2) refers to

Done. It refers to section 2. Rephrased to: " see Section 2".

214]"form" --> "from"

Done.

216] "For THE tag-and-probe"

Done.

223-226] "Impact...on differential cross sections" varies from 4 to 27%. Is this for the individual S, B components, or the ratio?

Done. As said, it is the overall impact, it includes simultaneous variations of S and B pdfs (we do all possible combinations and take the maximum values). Note that it is largely dominated by the uncertainty on the backgroud pdf.

228] range

Done.

228] Do 0.01 and 0.03-0.05 mean 1% and 3-5%?

No it doesn't mean %, those are actual variations of the pdf first bin. In percentage it would be ~1% for signal as the first bin is ~0.9, and ~10% for bkg as the first bin is ~0.3.

229] uncertainty on the

Done.

230] photon pseudorapidity $\eta$[,] number of vertices -> number of primary vertices $n_v$ -> $n_{vtx}$

Done.

231] $\eta$[,] and

Done.

237]has been assigned to the LHC integrated luminosity [16] -> has been assigned to the integrated luminosity corresponding to the dataset [16]

Done.

250-1] here state the ET1, ET2 range, and ii's eta < 2.5 excluding the 1.44-1.57 band, right? Applies also to words on figs.

Done

250] List barrel region (|eta|<1.44) first to be consistent with other usage. Same comment for 270.

Done.

257] P0[20][,] and DWT

Done.

258] (4.7 +/- 0.3) % ... can it really be so precisely known?

Indeed the 0.3% were just the statistical error on the computation of this correction. We have corrected it by the systematic error on this correction factor which has been estimated by considering different tunes. Text changed to "The parton-level cross section is corrected by a factor $0.95 \pm 0.04$."

265] factorisation[,] and fragmentation

Done.

267] lower -> less

Done.

274] Fig. 2 to 9 -> Figs. 2 to 9

Done. Changed to Figures 2 through 9.

274]Tables 2-5 are very nice to see.

275] What are the "four measured differential spectra"? Is it the diphoton mass, pt, delta-phi and costheta*? Just say so.

It seems to us unnecessarily to repeat here the list of considered observables.

276] Figs 5,6, and 7

Now both |eta|<1.44 and |eta|<2.5 figures are referenced.

281] "shoulder in the vicinity of 40 GeV". In Fig. 4 I observe a discrepancy in the whole pt > 40 GeV region. Do the authors mean the high-pt region (above 40 GeV) or just around 40 GeV?

We mean specially at 40GeV

287] Analogous --> Similar

Done.

288] Ref.[27].

Done.

294] azimuthal angle difference[,] and

Done.

301-303] There is not satisfactory agreement, eg Fig 7a is bad agreement over most of dphi.

Done. Text changed to: "Whereas there is an overall agreement between data and theory for the mass spectrum, the theoretical cross section appears underestimated for regions of the phase space where the two photons are not collinear."

388] Now arXiv:1106.5131 subm to PRL

Done.

Figures:

1. Figure 1 : Use variable bins for fit projections

The plots is represented with the same binning than the one used in the fit. Using variable bin is indeed a good suggestion. Nervertheless we don't think that the improvement, that it will provide, justifies a change of the figure as this stage of the review, while the figure has already been approved and made public. In addition it might be argued that using a different binning that the one used in the fit might be missleading. So we prefer not modify the figure.

2. Figure 1: What are $\gamma1$ and $\gamma2$?; Fig 1: ecallso ??; Example of the projections..Why is this an "example"? What is "ecallso"?

Figure legend and caption have been improved. New version should clarify these questions.

3. Figures [2-9]a: In general the "luminosity error" is not visible in the plots. Delete the "in quadrature" symbol.

Uncertainties from the luminosity measurement have been removed from the plot and indicated in the figure caption and in the text.

4. Include the name of the theory in the plots

Diphox and Gamma2MC indicated in the differential cross-section plots.

Comments from Albert de Roeck, 12 Aug 2011

Dear Editors

Thanks for your paper! A nice peace of work

The main visible result is the discrepancy observed w.r.t. theory and of course begs the question whether we understand the data well enoguh. At this time I judge our analysis to be solid solid and found no obvious issue, but I trust the ARC& auhtors did not take this lightly and looked in all corners if there could be problems with the data??

General comments

The results discussion: Not sure how to interpret the expression 'fair agreement' here in diplomatic language but clearly there is no real good agreement between TH and the experimental results. In fact we see important disagreements in all but perhaps the di-photon mass plot. Your explanation seems plausible, so the real conclusion of the paper is that we need improved calculations, right? (or would the conclusion be that we put ourselves into trouble by our phase space selection, which does not match well with the available calculations, ie we could have avoided that?) But I believe our written conclusion we say that there is an overall satisfactory agreement, which is in contrast what we show, and I believe this needs to be rephrased properly. We shouldn't be afraid of telling that the present theory implementation is not insufficient.

Done. Comments on theory/exp comparison were changed

We do not refer to ATLAS as far as we can see. Bad bad!! And CMS will get complaints in management from ATLAS. So please add upfront a reference to their paper released on July 4th arXiv:1107.0581.

Done. Added upfront.

They also see the disagreement in Deltaphi like we, but seem to have less issues for the other distributions. BTW I liked their introduction a lot, saying upfront that the QCD models they are about to use have missing ingredients, like soft gluon emission which is not in the fixed order calculations. Do we discuss the fragmentation contribution at all?

Fragmention was mentionned in the description of included processes. We have added mention of the fragmentation when discussing dphi spectra and inclusion of only one effective order.

Also I noted the ATLAS (trigger & Selection) thresholds are much lower than ours for the photons (15 GeV/selection offline of 16 GeV) Do we understand where that difference comes form? Are they really much better on the photon triggers compared to us?

We could actually lower the cuts to 18~19GeV, as our trigger with one isolated photon had a threshold of 17GeV. Trigger threshold could have been lowered by requiring a more stringent isolation criteria or by requiring both photons to be isolated. Although ATLAS does not mention it in their paper, they must probably require isolation in the trigger selection. For this measurement ATLAS trigger was better than ours, most probably due to different choices in the trigger selection.

What I am totally missing in this paper are basic experimental plots that show that we control the selections, purity, kinematics of the objects etc. Again this is done better in the ATLAS paper: it gives the reader more confidence that we are in control. I'd like the ARC & authors to think about that and perhaps ad a few key plots.

After having considered with the ARC to add basic experimental plots and look at the different plots that could be added, we have decided to just add a figure with the signal and background PDFs.

Since the comparison with theory is remarkable (and shows issues) I'd expect more details on the models used, like which version, critical parameters settings (alpha_s?) and other relevant parameters. But we definitely have the give at least the versions/tunes etc we use.

Done. Indicated alpha_s value used and version of the codes.

Details

- line 4: issuing? Is this the right word for the purpose here?

Done

- line 50 eta< 1.56 and line 62 eta< 1.57. Please be consistent

Done

- line 58: what trigger thresholds were used? suggest to add this information in the paper

Done. Thresholds has been varied during the 2010 data taking period. Maximum value of the HLT Pt threshold has been added.

- line 101 epsilon^sim -> epsilon^data?

Done

- line 177-182: would it not be worthwhile to show that with a figure? A priori it is not clear that we are so insensitive to pile-up, as found in this analysis, and could have been a concern for the reader.

We do not have approved plot showing this dependency. Since, as explained in this paragraph, the effect is taken into account, we don't think such plot is absolutly necessary.

- line 187: is there an ordering in what you call photon 1 and photon 2 in the event or is it just random? (in other words should fig 1 a and b be the same up to statistical fluctuations?)

Done: added: "the distributions of the ECAL isolation variables of the two photons, $\mathcal{I}^{1}$ and $\mathcal{I}^{2}, where numbers $1$ and $2$ are assigned randomly."

- line 202: can you given an indication of the size of this bias? worth to mention in the paper?

Done. Added: "It is less than 10% of the statistical error in 80% of the bins and never exceeds half the statistical error."

- line 204: Presumably the effect is small but did you check whether the unfolding result depends on the used MC to extract the matrix? Is there a systematic for this potential dependence? Line 213 indicates that migration effects can be as large as 40% at edges? I don't see anything in your systematics table related to the unfolding.

This dependency was checked by calculating the unfolding both with an MC sample containing only the "box" process and with an MC sample containing only the "born" process and was found to be negligible.

- line 258: perhaps this is my misunderstanding, but here we reduce the parton level cross section by 4.7% for UE and hadronization effects. However, are we not correcting our data for this loss? Ie do we really have to reduce the patron cross sections? Maybe this need a bit better clarification at this point in the text.

We correct our data back to the generated level with UE/hadronization by the unfolding procedure. The theoretical predictions are evaluated at parton level and need to get corrected for UE/hadronization effects to be compared to the unfolded data.

- PDFs: again maybe my understanding, due to the way it is written, but you say you take the half width of the envelope as uncertainty. The uncertainty is of course the full envelope but I guess you mean you take half the value to quote the +/- uncertainty range?? Or did you do something else, ie take a smaller uncertainty that you should when following the prescription..

Done. Fixed with: "The error on the cross section is taken to be the envelope."

Good Luck

cheers, Albert

Comments from Teruki Kamon, 11 Aug 2011

Congratulations for authors to complete a set of important measurements in 4 spectra (mass, pT, phi and cos(theta). Especally, cos(theta) is not reported by ATLAS (http://arxiv.org/abs/1107.0581).

The paper is well written - we read the text very smoothly. Below are our Type B and Type A comments. Hopefully, they are helpful.

Sincerely, Teruki Kamon for Texas A&M Universiy group

Type B

(*) Table 1

Why you used the symbol "~" for "differential cross-section uncertainty"? You showed nice differential cross-sections in Figs.2-9. Should we show the ranges?

The table intends to give an idea of how large are the systematic uncertainties, but the dsigma/dX uncertainties actually depend on the X bin. Few bins have a much larger systematics than the average, so from the systematics range we do not get the idea of how large is the systematics in most of the bins. It is why we provided a single approximate value, which was calculated as a weighted mean of the systematics, using the dsigma/dX as weights.

The value of the total systematics for each bin is indeed provided on the Figures and in Tables 2 to 5.

Total cross-section columns were removed. The \sim symbols were dropped and replaced with one line of an explanation in the caption.

(*) L. 275-276 ... in Fig.6, the theoretical cross section for dPhi < 2.8 is underestimated... ==> we see the data and theory in the last bins are not in agreement, either. So this sentence is misleading. Can we say "The theory doesn't predic the shape"? Same comment on Fig. 7

Done

(*) Paragraph starting L. 275 ... are in fair agreement with four measured differential spectra.... ==> We think the word "fair" is not well defined. Some readers may disagree. Could you re-write this paragraph with more specific? "... agree within the uncertainty in range X to Y", for example?

Done

(*) The most key aspect of this analysis is how well you understand the isolation shapes for signal and BG. The uncertainties are specified, but it would be useful to show the distributions for signal and BG separately with the uncertainty bands? See, for example, Figs. 4 of http://arxiv.org/abs/1107.0581

The distribution for signal and background photons of ECAL isolation variable has been added to the paper (Fig. 1 on the new draft).

Type A

(*) abstract The word "satisfatory" depends on how people feel. So "overall agreement" is good enough.

Done

(*) L.43-44 .. are the **magnitudes" of the transverse momenta ..

Done

(*) L.96, L.98, L.101 Add "comma" at the end of sentence. (L.96, L.98) Add "period" at the end of sentence (L101).

Done

(*) L.141 ... 5 crystal wide strip alonh]g phi are removed. ... You mixed "belt" (L. 79) and "strip" in the text. Please unify.

Done: "strip" chosen

(*) L.175 and (L.302 Please avoid using "satisfactory". You can re-phrase by saying "The agreement is within 1 sigma of uncertainty in entire range of the {\cal I} distribution." for example

Done .

(*) Fig. 1 The x-axis labels "ecaliso" should be {\cal I} (L.132).

Done

(*) L.221, L.262, Figs. 2,3, 4,5,6,7,8,9 "errors" ==> "uncertainties"

Done

(*) L. 274 Fig. 2 to 9 ==> Figs. 2 through 9

Done

(*) L.277, L278, L.279 "LO" and "NLO" are not defined.

Done

(*) L. 297 "... electromagnetic energy isolatoion ...."

Done

(*) Refs. 1, 6, 7, 14 The page number range is shown, but not in other references.

Done

(*) Refs. 6, 27 (Oct, 1977) ==> (1977) (May, 2001) ==> (2001)

Done

(*) Ref. 18 What is "MP108"? A session number? Needed?

.. MP108, October 27-31, 2008 (2009) 12-31 .. This is mixing the conferencve dates and proceeding information. Please fix.

Done

(*) Ref. 20, 26 Space between "Phys." and "Rev." Space between "Phys." and "Lett."

Done

(*) Ref. 25 CDF Public Note ... ("Public" is added).

Done

Comments from Sijin Qian, 9 Aug 2011

I have roughly read through the QCD-10-035-paper-v0.pdf, and would have some (large and small) questions and comments from a non-expert's point of view. I list them below, please make a note of it if any of them would be sound.

Thank you and looking forward to hearing from you,

-Sijin

---

begin --------------------------

In general

(1) Throughout the paper (including the table captions, etc.), the subscript "T" in most of variable "pT"s apparently are placed a little too high, like in the middle part of total height of "p". If the bottom of "T" is placed at the similar level of the bottom of "p", it will be looked much better.

Done using the CMS predefined \pt, \ET, \kt

(2) In all figures, on the line above the plots, the word of "preliminary" should be removed.

Done

(3) Throughout the paper, I'm not sure whether or not the "integrated cross section" should be changed to the "total cross section".

It has been decided to use the 'integrate cross section' term during the review within the QCD PAG. The cross section being computed for a limited acceptance, this terms has turned out to be more appropriate than 'total cross section'.

(4) On L50, it is printed that the endcap ECAL starts from |eta|>1.57; while on L62, it is printed that the endcap ECAL fiducial region is |eta|>1.56. Afterward, these two numbers (i.e. 1.56 and 1.57) are used alternatively: e.g. on L127 and in Table 1, it is 1.56; while on L250 and in Tables 2-5 and Figs. 2, 4, 6, 8's captions, it is 1.57. I wonder whether they all should be consistent.

Done, all is 1.57 now

Page 0

(5) In the title of this paper, the 2nd line, it seems slightly better to insert a space before

and after the "=" sign respectively, i.e. "in pp collisions at sqrt(s)=7 TeV" --> "in pp collisions at sqrt(s) = 7 TeV"

Done

Page 1

(6) L2-4, there are 3 "of"s and 2 "collid"s close each other at the beginning of sentence, it seems that the sentence may be re-arranged and shortened from "The study of the production of pairs of energetic photons in hadronic collisions at collider energies is a valuable testing ground of the perturbative regime of Quantum Chromodynamics (pQCD)." --> "The study of the pair production of energetic photons on hadron colliders is a valuable testing ground of the perturbative Quantum Chromodynamics (pQCD)."

Done. Rephrased to: "The study of the production of energetic photon pairs in hadronic collisions is a valuable testing ground of the perturbative ..."

(7) L4, I'm not sure sure whether the word "issuing" should be changed from "Prompt diphotons issuing from hard parton-parton scatterings" --> "Prompt diphotons issued from hard parton-parton scatterings"

Done. Rephrased to: "The emission from hard parton-parton scattering of a pair of photons constitutes..."

(8) L8, there are 2 "for"s close each other, the 1st one may be changed from "For searches for new particles," -->"Regarding to searches for new particles,"

Done. Rephrased to: "In searches for new particles, ..."

(9) L13, similar as the item (6) above, one of 3 "of"s can be reduced from "a measurement of the production cross section of pairs of isolated photons" --> "a measurement of the production cross section of isolated photon pairs"

Done. Rephrased to: "a measurement of the production cross section of isolated photon pairs"

(10) L18, I'm not sure whether the uncertainty of luminosity should be mentioned of not, i.e. "with an integrated luminosity of 36 pb-1" --> "with an integrated luminosity of 36 +- 1.4 pb-1 [16]"

Done: uncertainty added

(11) L26-27, because (a) only polar (but not azimuthal) angle is w.r.t. the beam axis; (b) the beams have two directions (clockwise and anti-clockwise); (c) the numerical value of the angle phi has been shown on L276 (and etc.) of Page 8, and an angle can be measured in either the radians or degrees, therefore, the unit of phi should be specified; thus, this sentence seems should be changed from "theta and phi respectively designate the polar and azimuthal angles with respect to the beam axis." --> "theta and phi respectively designate the polar (with respect to the anti-clockwise beam axis) and azimuthal (in radians) angles."

Done. Rephrased to: "In the CMS coordinate system, $\vartheta$ and $\varphi$ respectively designate he polar angle with respect to the counterclockwise beam direction and the azimuthal angle, expressed in radians throughout this paper."

(12) L39, two words may be added as "with a total of 3 X0 of lead" -->"with a total radiation length of 3 X0 of lead"

Done

(13) L41-46, this sentence is very long with many commas, it seems much better and clearer if it can be separated by several bullets ended with semi-colons, and to add a word of "where" at the end of the line before the last line, e.g. "The differential cross section has been measured as a function of the diphoton invariant mass, m(gg), the azimuthal angle between the two photons, Delta(phi(gg)), the transverse momentum of the photon pair, pT, gg = sqrt(...), where pT,g1 and pT,g2 are the transverse momenta of the two photons, and the absolute value of the cosine of the angle between the two photons in the Collins-Soper reference frame [6], cos(theta*)= tanh(Delta(ygg))/2, Delta(ygg) being the difference between the two photon rapidities." --> "The differential cross section has been measured as a function of (a) the diphoton invariant mass, m(gg); (b) the azimuthal angle between the two photons, Delta(phi(gg)); (c) the transverse momentum of the photon pair, pT, gg = sqrt(...), where pT,g1 and pT,g2 are the transverse momenta of the two photons; and (d) the absolute value of the cosine of the angle between the two photons in the Collins-Soper reference frame [6], cos(theta*)= tanh(Delta(ygg))/2, where Delta(ygg) being the difference between the two photon rapidities."

Done: add list with bullets

Page 2

(14) L49, the variable "R" seems should be defined or referred here at its first appearance, i.e. "by R = 0.45 within ..." --> "by R = ... (a definition expression) = 0.45 within ..." or "by R = 0.45 (where R is defined in [xx]) within ..."

R is actually defined on line 28

(15) Between L54 and L55, at the end of Introduction Section, normally it has a paragraph to briefly introduce each of other Sections in the paper, e.g. "In this paper, Section 2 is ...; Section 3 ...; ... are described in Section 4; ...; the conclusions are in Section 8". But this paper seems missing this paragraph yet.

DONE

(16) L79-82, this sentence is very long with a long explanation in the middle being sandwiched by the subject ("no charged particle") and its verb ("impinges") of this sentence. It seems much better and clearer if the long explanation can be put into a pair of bracket, or be re-arranged, i.e. "it is required that no charged particle with pT > 3 GeV, with impact parameters in the transverse and longitudinal planes of less than 1mm and 2mm respectively, and associated with a hit in the innermost layer of the pixel detector, impinges on ECAL within a cone of radius R = 0.4." --> "it is required that no charged particle (with pT > 3 GeV, the impact parameters in the transverse and longitudinal planes of less than 1mm and 2mm respectively, and associated with a hit in the innermost layer of the pixel detector) impinges on ECAL within a cone of radius R = 0.4." or "it is required that no charged particle impinges on ECAL within a cone of radius R = 0.4; here the charged particle refers to the ones with pT > 3 GeV, the impact parameters in the transverse and longitudinal planes of less than 1mm and 2mm respectively, and associated with a hit in the innermost layer of the pixel detector."

Rephrased.

(17) L86-87, the NLO has been used later in the paper, but has not been explained yet. It should be done at its first appearance here, i.e. "using the next-to-leading order POWHEG generator [10?12]." --> "using the next-to-leading-order (NLO) POWHEG generator [10?12]."

Done

Page 3

(18) L98, a comma seems missing, i.e. "in Section 1" --> "in Section 1,"

Done

(19) L99, it seems that a word should be added as "is the efficiency of the identification criteria measured from data," --> "is the efficiency of the photon identification criteria measured from data,"

Done

(20) L100-101, it seems (a) the last term in the sentence has an error; (b) a period is missing at the end. Namely, it should be changed from "epsilon(sim) is the efficiency ... using the same technique as for epsilon(sim)" --> "epsilon(sim) is the efficiency ... using the same technique as for epsilon(data)."

Done

(21) L102-103, 3 letters of "di-" seem missing, i.e. "The efficiencies epsilon(data) and epsilon(sim) to observe a photon candidate are taken as the square of the efficiencies to observe a single photon." --> "The efficiencies epsilon(data) and epsilon(sim) to observe a di-photon candidate are taken as the square of the efficiencies to observe a single photon."

Done

(22) L107, I'm not sure whether it would be slightly better to change from "while the other, the probe, is selected" --> "while the another, the probe, is selected"

Don't agree

(23) L119-120, it seems should be changed from "The cone is also required not to include photon and electron candidates, nor jets." --> "The cone is also required to neither include photon and electron candidates, nor jets."

Done. Rephrased to: "The cone is also required not to include photon and electron candidates or jets."

Page 4

(24) L139, the spaces before and after the ">" sign may be reduced a bit, i.e. "ET > 300 MeV" --> "ET > 300 MeV"

Done

(25) L152-153, the "two additional independent methods defined in Section 2" has been mentioned,

it seems better (for readers) if the names of two methods can be reminded here, i.e. "two additional independent methods, both exploiting low-radiating e+ and e-, defined in Section 2, ..." --> "two additional independent methods (i.e. name of method 1 and name of method 2), both exploiting low-radiating e+ and e-, defined in Section 2, ..."

Done: comma removed, "defined in section 2" is referring to "low radiating electrons", not to the methods which are described thereafter

Page 5

(26) L203-204, I have not yet understood the "(2)" at the end of the sentence "The number of signal events estimated with the fit has been corrected for the reconstruction and identification efficiencies (2)." (a) Would it refer to Eq.(2)? But Eq.(2) is the likelihood function, but for the correction? (b) If the answer to (a) is "no", then would it should be something else?

Done. It refers to the section, rephrased to: "see Section 2"

Page 6

(27) In the caption of Fig.1, (a) on the 2nd line, it seems slightly better to change from "The continuous blue line" --> "The solid blue line"

Done

(b) on the 3rd line, the "dashed red line" can not be clearly seen in the black-white print. It will be appreciated if some improvements can be made on this line.

Done: plot changed

(28) L212 and L214, I can not clearly figure out what the "former" and "latter" in these two sentence are referring to. It seems much better and clearer (for readers) if the explicit names can be used instead of the "former" and "latter".

Done. Rephrased to: "This affects the definition of the acceptance and induces bin-to-bin migrations in the differential cross sections. The former impacts only kinematics regions near the photon pT thresholds and results in an uncertainty of 40% in the most affected region, the lowest masses of ds/dm. The uncertainty from the bin-to-bin migration} is of about 1%."

(29) L214, I wonder whether there is a typo, i.e. "uncertainty form the latter is of about 1%." --> "uncertainty from the latter is of about 1%.

Done

Page 7

(30) Table 1

(a) In the 2nd sentence of the caption, it may sound better to re-arrange it and a letter "s" may should be added at the end of the word "contribution", i.e. "In this table are listed the different sources of systematic uncertainties on the measured cross section with their respective contribution." --> "In this table the different sources of systematic uncertainties on the measured cross section with their respective contributions are listed."

Don't agree

(b) On the 4th line of the header row, it seems should put no space before and after the 3 signs of "<", i.e. "|eta| < 1.44 or 1.56 < |eta| < 2.5" --> "|eta|<1.44 or 1.56<|eta|<2.5"

Done

(c) In the data cells, all percentages seem should have the same number of digit after the decimal point, i.e.

in the right-most column:

the 1st data row "~2%" --> "~2.0%"

the 3rd data row "~9%" --> "~9.0%"

the 4th data row "~3%" --> "~3.0%"

the bottom "Total" row "~11%" --> "~11.0%"

in the middle-right column

the 2nd data row "~1%" --> "~1.0%"

the 3rd data row "~7%" --> "~7.0%"

the 4th data row "~2%" --> "~2.0%"

the bottom "Total" row "~8%" --> "~8.0%"

The number of digit has been chosen according to their significance.

(31) L240, the last word seems should not be capitalized, i.e. "as follows: Contributions" --> "as follows: contributions"

Done

(32) L248, it seems very rarely to use the "gluon luminosity" in the literature. Is it possible to choose another popular words?

"gluon luminosity" is used for instance in Eur. Phys. C16(2000) 311-330, T. Binoth et al.. We are open to other proposal.

(33) The line (without line number) below L258, similar as the item (27) above, I can not clearly figure out what the "latter correction" in the sentence is referring to. It seems much better and clearer (for readers) if the explicit name of correction can be used instead of the "latter correction".

Part rephrased.

(34) Two lines (without line numbers) above L259, (a) as the "parton distribution function" has been abbreviated as "PDF" on L260, I'm not sure whether it should be changed from "2.0 (pdf) pb, 1.1 (pdf) pb." --> "2.0 (PDF) pb, 1.1 (PDF) pb."

Done

(b) the 1st uncertainties on both lines are for "(scale)", is it possible to very briefly explain what the "scale" in this Section is about?

Done: "scale" -> "scales" as in the text

Page 8

(35) The line (without line number) below L270, I'm not sure whether there is a typo or not, i.e. "The integrated cross sections obtained for the acceptance defined in Section 2 are:" --> "The integrated cross sections obtained from the acceptance defined in Section 2 are:"

Don't agree

(36) L274, a letter of "s" is missing, i.e. "in Fig. 2 to 9." --> "in Figs. 2 to 9."

Done

(37) L277, the term of "LO" should be explained at its first appearance here, i.e. "In the LO term of gluon fusion ..." --> "In the Leading Order (LO) term of gluon fusion ..."

Done

Page 9

(38) Fig.3's caption seems can be shortened from "Figure 3: Measured cross section of diphoton production as a function of the invariant mass of the photon pair (a) and bin-by-bin comparison with the theory (b) for photons within the pseudorapidity region |eta| < 1.44. The total systematic uncertainties are represented by the shaded area, the different contributions are added in quadrature sequentially." --> "Figure 3: The same as Fig.2, but for photons within |eta| < 1.44."

We prefer to keep the caption independent.

(39) Fig.4's caption can be similarly (as the item (38) above) shortened from "Figure 4: Measured cross section of diphoton production as a function of the transverse momentum of the photon pair (a) and bin-by-bin comparison with the theory (b) for photons within the pseudorapidity region |eta| < 1.44 or 1.57 < |eta| < 2.5. The total systematic uncertainties are represented by the shaded area, the different contributions are added in quadrature sequentially." --> "Figure 4: The same as Fig.2, but as a function of the transverse momentum."

Same as previous comment.

Page 10

(40) Fig.5's caption can be similarly (as the item (38) above) shortened from "Figure 5: Measured cross section of diphoton production as a function of the transverse momentum of the photon pair (a) and bin-by-bin comparison with the theory (b) for photons within the pseudorapidity region |eta| < 1.44. The total systematic uncertainties are represented by the shaded area, the different contributions are added in quadrature sequentially" --> "Figure 5: The same as Fig.4, but for photons within |eta| < 1.44."

Same as previous comments.

(41) Fig.6's caption

(a) can be similarly (as the item (39) above) shortened from "Figure 6: Measured cross section of diphoton production as a function of the azimuthal angle between the two photons (a) and bin-by-bin comparison with the theory (b) for photons within the pseudorapidity region |eta| < 1.44 or 1.57 < |eta| < 2.5. The total systematic uncertainties are represented by the shaded area, the different contributions are added in quadrature sequentially." --> "Figure 6: The same as Fig.2, but as a function of the azimuthal angle between the two photons."

Same as previous comments.

(b) The axis variable of "delta(phi(gamma-gamma))" should have a unit for each plot, i.e. "delta(phi(gamma-gamma))" --> "delta(phi(gamma-gamma)) (radians)"

It has been added in the introduction that all angles in the paper are express in radians.

Page 11

(42) Fig.7's caption

(a) can be similarly (as the items (38) and (40) above) shortened from "Figure 7: Measured cross section of diphoton production as a function of the azimuthal angle between the two photons (a) and bin-by-bin comparison with the theory (b) for photons within the pseudorapidity region |eta| < 1.44. The total systematic uncertainties are represented by the shaded area, the different contributions are added in quadrature sequentially." --> "Figure 7: The same as Fig.6, but for photons within |eta| < 1.44."

Same as the similar previous comments.

(b) Similar as the item (41b) above, the axis variable of "delta(phi(gamma-gamma))" should have a unit for each plot, i.e. "delta(phi(gamma-gamma))" --> "delta(phi(gamma-gamma)) (radians)"

Same as the similar previous comment.

(43) Fig.8's caption can be similarly (as the items (39) and (41a) above) shortened from "Figure 8: Measured cross section of diphoton production as a function of the absolute value of the cosine of the polar angle between the two photons in the Collins-Soper reference frame and of the photon pair (a) and bin-by-bin comparison with the theory (b) for photons within the pseudorapidity region |eta| < 1.44 or 1.57 < |eta| < 2.5. The total systematic uncertainties are represented by the shaded area, the different contributions are added in quadrature sequentially." --> "Figure 8: The same as Fig.2, but as a function of the absolute value of the cosine of the polar angle between the two photons in the Collins-Soper reference frame and of the photon pair."

Same as the similar previous comments.

Page 12

(44) Fig.9's caption can be similarly (as the items (38), (40) and (42a) above) shortened from "Figure 9: Measured cross section of diphoton production as a function of the absolute value of the cosine of the polar angle between the two photons in the Collins-Soper reference frame and of the photon pair (a) and bin-by-bin comparison with the theory (b) for photons within the pseudorapidity region |eta| < 1.44. The total systematic uncertainties are represented by the shaded area, the different contributions are added in quadrature sequentially." --> "Figure 9: The same as Fig.8, but for photons within |eta| < 1.44."

Same as the similar previous comments.

(45) L280, it seems better to add a word of "and", i.e. "with the requirements of ET > 20, 23 GeV" --> "with the requirements of ET > 20 and 23 GeV"

We prefer the original text. We will follow the PubCom recommandations if they think the text need modifications.

Pages 13 and 14

(46) In Tables 2-5, on the 3rd header rows of all 4 Tables, in the right columns, more spaces seems should put before and after the word of "or", i.e. "|eta| < 1.44 or 1.56 < |eta| < 2.5" --> "|eta| < 1.44 or 1.56 < |eta| < 2.5"

Done

Page 14

(47) Table 4, (a) on the 1st header row, the unit should be changed from "[pb]" --> "[pb/radians]"

It has been added in the introduction that all angles in the paper are express in radians.

(b) on the 2nd header row, the left column, the unit should be added, i.e. "Bin" --> "Bin [radians]"

same as previous comment.

Page 15

(48) L290, one of two successive "of"s may be replaced as "production cross sections of pairs of isolated photons have been" --> "production cross sections of isolated photon pairs have been"

Done

(49) L301, would it be possible to very briefly explain what the "fixed order computation" is?

It is described in the Theoretical Prediction section. We have added citations of the reference paper for Diphox and Gamma2MC.

(50) L304, according to the common practice in all CMS papers, the Acknowledgment Section does not need a Section series number, i.e. "9 Acknowledgments" --> "Acknowledgments"

Done

Pages 16-17, in the References Section,

(51) L327, in [1], according to the common practices in all HEP papers, if the number of authors is more than 3, only the 1st author is put in the Reference plus "et al.", instead of "2 or 3 names + et al.", i.e. "[1] T. Binoth, J. P. Guillet, E. Pilon et al.," --> "[1] T. Binoth et al.," The other ones which need to be similarly modified are Refs. [4], [10], [23] and [27].

Done

(52) For the consistency in this Section, all references should have only one page number instead of two. The ones which need to be modified are Refs. [1], [6], [7], [14], [23] and [26].

Done

(53) L334, in [3], the words in the journal title should be separated by a space, i.e. "Phys.Rev. D80 (2009) 114016" --> "Phys. Rev. D80 (2009) 114016" The other ones which need to be similarly modified are Refs. [20] and [26].

Done

(54) L342, in [6], to be consistent with other PRD references (e.g. [2]-[4] and [22], etc.), (a) only the year number but no month should be given to each reference, (b) there should be no space between "D" and volume number "16", and "D" should be non-italic and bold, i.e. "Phys. Rev. D 16 (Oct, 1977) 2219?2225." --> "Phys. Rev. D(non-italic and bold)16 (1977) 2219." The other one which needs to be similarly modified is Ref. [27].

Done

(55) L349, in [9], a pair of bracket is missing, i.e. "Isolated Photon Reconstruction and Identification at sqrts = 7 TeV" --> "Isolated Photon Reconstruction and Identification at sqrt(s) = 7 TeV" The other one which needs to be similarly modified is Ref. [13]. But strangely, [9] and [13] have the identical PAS number, but with the different titles.

Done

(56) L389, in [25], it should be changed from "with 5.4 fb-1 of CDF Run II Data", CDF note 10160 (May, 2010)." --> "with 5.4 fb(superscript '-1') of CDF Run II Data", CDF note 10160 (2010)." Please notice that there should be no "month" before the year number as mentioned in the item (54a) above.

Done

Comments from Eiko, 8 Aug 2011

Dear Authors and the ARC members,

Congratulations on the first CMS measurement of the diphoton cross section!

Below you could find my comments for the draft.

Abstract: "photons radiated at small relative angle" ---> I assume you mean relative angle between the two photons, but the text may be interpreted as the angle between photon and the parton that emits photon. Suggest to change it to "populated by photon pairs with small angle difference between the two photons" or something similar.

Done.

L15-17 Photons may be prompt, but not isolated. Isolated photons may not come from promp photons and may come from decays of hadrons (when there's only one leading pi0 in a jet). Need to re-word these three lines, maybe define "signal photon" and "background photon"?

We have now defined what is called "signal photon" and "background photon" at the beginning of the paper and we refer to these terms throughout the paper

L13-18 Need to mention briefly and cite ATLAS 37/pb result. http://inspirebeta.net/record/916832

Done.

L56-59 What are the trigger requirements? It's not clear from these lines that there are ECAL isolation cuts and that one needs to apply ECAL isolation cut offline to be consistent with trigger cuts (as described in L76--77).

Thresholds has been varied during the 2010 data taking period. Maximum value of the HLT Pt threshold has been added. Indication of the isolation requirement in the trigger has been added.

L67 and through out the paper: sη ---> σηη

Done.

L73 −−−>\Delta \eta = 0.015$

Corrected. 0.015 is actually the half-width. \Delta \eta = 0.03

Between L93 and L94, equation (1): why do you call A(X_i) an acceptance correction factor, rather than the product of acceptance and efficiency? On top of which variable is A(Xi) correcting?

To me, acceptance (actually efficiency) correction factor refers to the data to simulation correction factor (the second ratio in Equation 1) only.

This part was reworked to make it more clear.

L95-98 There's no description about simulated samples. I assume you meant that N_RECO^SIM and N_GEN^SIM are obtained from simulated diphoton events. Would be nice to mention how they were generated. Are they the same as what is described in the first paragraph of Section 6?

Done.

L102 ---> to observe a diphoton candidate

Done.

L126-127 --> mention that the uncertainties on the acceptance correction factor will be discussed in Section 5.

I don't think it is necessary, while forward references make the reading unpleasant, asking the reader to go back and forth. Shall we add it?

L130 hadron decays --> decays of neutral hadrons

This part was rephrased.

L167-L171 Do we know the estimated number of impinging tracks from MC background? I remember Laurent showed a few studies of that. It's worth mentioning why you focus on 1-impinging-track events for the central background shape. Is this going to bias your background isolation towards smaller value?

The signal selection requires that there is no impinging track. So the bias is actually expected to larger value of the isolation variable. The energy deposited around the track is excluded from the isolation sum in order to avoid this bias. Uncertainties coming from this possible biais is estimated by the procedure explained L172-175 (draft v0).

The normalized histogram of the number of impingings for signal and background is shown below.

L173 sorry I forgot, how did you choose one of the two tracks to be excluded in the computation of I? Randomly?

It was indeed choosen randomly.

L175 "the track is in this case left uncorrected.": which track? uncorrected for what?

The energy deposited around the track is not excluded from the isolation sum. Text rephrased to make it clearer.

L201 toy Monte Carlo experiments --> pseudo-experiments

Done.

L204 identification efficiencies (2) --> identification efficiencies as described in Section 2.

Was changed to "see Sec. 2".

L205 Cite the method you used to perform unfolding?

Done

L210 Reference [15] should be changed to the following and 1.4% should be changed to 1.5%.

Done.

@ARTICLE{ref:EGM-11-001, AUTHOR = "{CMS Collaboration}", COLLABORATION = {CMS}, TITLE = "ECAL 2010 performance results", URL = "http://cdsweb.cern.ch/record/1373389", JOURNAL = "CMS Detector Performance Summary", VOLUME = "\href{http://cdsweb.cern.ch/record/1373389}{CMS-DPS-2011-008}", YEAR = "2011" }

Done. But are we expected to cite a DPS?

L288 Need to replace Ref [25] with the following one:

@Article{Aaltonen:2011vi, author = "Aaltonen, T. and others", collaboration = "CDF", title = "{Measurement of the Cross Section for Prompt Isolated Diphoton Production in p\bar p Collisions at \sqrt{s} =

1. 1. 1. 1. 1. 1. 96 TeV}", year = "2011", eprint = "1106.5123", archivePrefix = "arXiv", primaryClass = "hep-ex", SLACcitation = "%%CITATION = 1106.5123;%%"

} Done.

Addendum from 9 Aug 2011

Dear Authors,

I was told that the CDF result is actually published in two papers (one in PRL and one in PRD). Both should be the same and the PRD just includes more details. Both are accepted just yesterday. It'd be nice to cite both papers.

The arXiv links are:

arXiv:1106.5123 PRL

arXiv:1106.5131 PRD

Done

Shin-Shan Eiko Yu

Comments from Steve Wasserbaech, 4 Aug 2011

Steven Wasserbaech (steven.wasserbaech@cernNOSPAMPLEASE.ch) Date of comment 04 Aug 2011 Note

My comments are in the attached pdf file. I erroneously posted them to the HN, for which I apologize!

Steve

Greetings. Here are some comments about QCD-10-035, v0.

> Type B
I agree that the paragraph on lines 41-52 belongs in the introduction. I also agree that the information on lines 26-32 needs to come before that. But I suspect lines 19-25 and 33-40 could and should appear later. They aren't really needed for the introduction. Perhaps they could become Section 2, along with lines 53-54.

Kept original text

Remove "preliminary" from the figures. Also, can we please separate (e.g., with commas) the other bits of information that appear above the figures (eta range, sqrt{s}, integrated luminosity)?

Done

> Type A

general comments:

1. "Collins--Soper" should have an en-dash -- instead of a hyphen -.

Kept original text

2. No hyphen in "nonprompt" or "subsample".

We have now defined what is called "signal photon" and "background photon" at the beginning of the paper and we refer to these terms throughout the paper

Changed sub-sample to subsample

3. We normally use present tense to describe our analyses, except in the conclusions. At least we shouldn't jump back and forth between present and past erratically.

Done

4. If you are going to use passive voice in past tense outside of the conclusions, use "was" instead of "has been".

Done

5. I think it's OK to refer to the eta-phi space as a plane. But it doesn't make sense to refer to a "cone of radius R" in a plane. We really mean a circle in the eta-phi space. And we could simply say "with R < 0.4" or something like that.

No change done. While R is a distance in(eta, phi) plane, R < R_0 still defines a cone in (r, eta, phi) space.

6. Figures 2-9 captions: to me it seems unnatural and confusing to write "Figure X: Information about the first plot (a) and about the second plot (b) for photons blah blah." I would write "(a) Information about the first plot and (b) about the second plot, for photons blah blah."

%DONE% Done.

specific comments:

title: Introduce spaces before and after the "=", or delete "sqrt{s} =" altogether.

Done

abstract:

first line: "cross sections for the production"

Kept original text

fourth line: hyphen between "leading" and "order", because "next-to-leading-order" is being used a modifier

Done

last line: "photons radiated at small relative angle" is unclear

Done while implementing Eiko's comments.

line 12: "enhance the sensitivity"

Done

lines 17-18: I think this sentence would be better if it were reorganized. The present form seems awkward. And we shouldn't refer to the dataset before mentioning any datasets. A proposal: "The data sample was collected in 2010 and corresponds to an integrated luminosity of 36 pb-1."

Done

line 27: use \frac{\theta}{2} or put \theta/2 in parentheses

Done

line 28: "$\eta$-$\phi$" or remove the space after the comma in "(\eta,\phi)"

Done

line 30: insert "with" before "pz"

Done

line 39: sorry about the confusion: there should be no \, in "$3X_0$"

Done

line 41: "has been" -> "is"

Done.

line 45: either remove the parentheses in the equation or remove the \frac (use / instead)

Done

lines 46-47: don't say both "in addition" and "also"

Done

line 49: surely we mean "> 0.45" rather than "= 0.45"

Done

line 49: delete "the pseudorapidity"

Done

line 51: "the photon" -> "each photon"

Done

line 53: "has been" -> "is"

Done.

lines 69-70: the transverse momentum of what?

Clarified. Added: "of charged particles [...] around the photon direction".

line 93+1: no colon

Done.

eqn 1: comma at the end of the equation

Done.

lines 94-101: peculiar structure, which seems unnecessary. We can just use normal paragraph format. In any case, use a semicolon after each item, but a full stop after the fourth one.

semicolons and period added.

line 102: "photon" -> "diphoton"

Done.

line 104: "has been" -> "is"

Done.

line 185: "in" -> "into"

Done.

line 189: comma between "t" and "and"

Done.

fig 1: In the horizontal axis labels, the "I" is indistinguishable from the "l".

Done.

caption: It would be nice to mention which variables are plotted.

Done.

line 258: no space between ")" and "%"

Done.

displayed equations after line 258+2: the spacing is quite unnatural

Done.

line 292: comma or "and" (but not both) after "2010"

Done.

line 294: comma after "difference"

Done.

Sincerely, Steve

Comments from Greg Landsberg, 4 Aug 2011

Dear Authors and the ARC members,

Congratulations on a measurement of the diphoton cross section with 2010 data and a nicely written paper! I have a few minor physics comments and a number of style suggestions geared toward reducing the use of jargon (data-driven, error, toy MC) in the paper and making the text more clear in places.

PHYSICS COMMENTS:

1) L48: suggest putting the explanation why asymmetric cuts were used for photon pTs (trigger, robustness of NLO calculations?).

Done.

2) LL50-52: the statement about the isolation in the sentence here contradicts actual definition of the isolation given in LL69-83. Later in the text you say that these isolation criteria are applied to theoretical calculations. Yet, first of all the statement in the introduction is confusing, and second why one would expect the theory and the experimental data to agree for quite different isolation criteria? For one, the isolation used in data is relative (20% of the photon energy and has complicated dependence on HCAL), whereas the theoretical calculations use fixed 5 GeV isolation. How do we know that this difference does not result in, e.g., the discrepancy at small dPhi, where the two photons start to be close to each other?

Text of LL50-52 part has been clarified. In the calculation, contribution at low delta-phi can indeed be increased by releasing the 5GeV isolation cut, which is as you have noted already loose compare to the experimental cut. But the increase is small compared to the discrepancy with the measurements. However modeling in the calculation of the experimental isolation is crude and we cannot exclude that is not due to the limit of this model.

3) L257: give PYTHIA version (6.4xx).

Done

4) Ref. [25]: our standing agreement with ATLAS is not to cite each other unpublished results; this should also apply to the unpublished Tevatron results. Suggest removing the reference from the text.

To be checked with pub committee. Tevatron paper has been since published and we expect ATLAS paper to be published by the time we send the final proof for publication.

STYLE COMMENTS:

0) General: the absolute majority (all ?) other CMS papers use symbol θ to denote the polar angle; suggest switching to this convention through the paper.

Done. \vartheta changed to \theta

All the figures should have CMS and not CMS Preliminary in the header.

Done.

All the computer codes (POWHEG, Gamma2MC, DIPHOX, GEANT4, PYTHIA) should be typeset in small caps.

Small caps Done.

1) Introduction,

L2: energetic photons (diphotons);

Done.

L3: delete "regime of";

Done.

L4: issuing -> emitted;

Done.

L8: gravitons in models with extra dimensions;

Done.

L14: centre-of-mass; Compact Muon Solenoid (CMS);

Done both.

L15: Large Hadron Collider (LHC);

Done.

LL16-17: delete the whole sentence about diphotons, as the definition has been moved to L2, before the first use of the term.

Definition was left there since it includes the prompt property. Diphoton term was removed from the introduction

LL43,44: put "where ... of the two photons" in parentheses.

Not changed

L49: R > 0.45, within.

Done

2) Event Selection,

L63: R > 0.45;

Done

L68: in Ref. [9];

Done

L80: add a comma before "respectively"; Eq. (1): end with a comma;

Done

L98: end with a comma;

Done

L101: end with a period;

Done

LL104-105: ``tag-and-probe'' (in quotation marks) and add a reference here, e.g. to our first W/Z cross section paper.

Quotes and ref. to W/Z paper added.

3) Signal and Background Discrimination,

L136: obtained from data, using the;

Done

L138: The isolation variable I is defined ...;

Not changed

L141: five-crystal-wide strip;

Done

L149: on -> using

Done

L155: The Z→e+e− events ... (can't start a sentence with a symbol);

Done

L166: determined from data;

Done

L170: The isolation I is then normalised ...;

Done

L176: taken as a systematic uncertainty;

Done

L180: add an Oxford comma before "and".

Done

4) Cross-Section Measurement,

L184: fit to the distributions;

Done

LL189-191: combine two sentences as "... for each type t, Ntot is the sum ... three event types, and ft ...";

Done

L194: checked with simulation;

Done

L195: rephrase so that you do not start the sentence with a number;

Done

L197: add a comma before "or";

Done

L201: with Monte Carlo (MC) pseudo-experiments.

Done

5) Systematic Uncertainties,

L214: form -> from; also it is unclear what "latter" is referring to? Bin-to-bin migrations? Please clarify.

Done

L216: errors -> uncertainties;

Done

L224: with MC pseudo-experiments;

Done

LL230,231: add a comma before "and" (twice);

Done

L237: different -> various.

Done

6) Theoretical Predictions,

L240: contributions (small letter); Equations above

Done

L259: pdf -> PDF; "(scale)" is not explained until later.

Done

L262: errors -> uncertainties (twice);

Done

L264: errors -> uncertainties (twice);

Done

L265: add a comma before "and".

Done

7) Results,

L277: LO has not been defined;

Done

L279: NLO has not been defined; should be defined on L87.

Done

8) Conclusions,

L294: add a comma before "and";

Done

L301: state-of-the-art fixed order computations.

Done

9) Acknowledgements,

L305: add a comma before "and".

Done

10) References. Ref [1]: "A full ..." (no need to capitalize "full");

Done

Ref. [5]: JINST {\bf 03};

Done

Ref. [6]: Phys. Rev. {\bf D16} (1977) 2219-2225;

Done

Ref. [9]: s=7 TeV;

Done

Ref. [13]: it's the same as Ref. [9] - please remove and change everything to Ref. [9];

Fixed

Ref. [17]: add the doi reference;

Done

Ref. [19], L2: add a comma before "except"; L3: CTEQ5L PDF. Remove "(2010)".

Done

Ref. [20]: Phys. Rev. (add space);

Done

Ref. [27]: Phys. Rev. {\bf D63} (2001) 114016 and add arXiv reference - hep-ph/0012191.

Done

11) Figures.

Fig. 1: use {\cal I} instead of "ecaliso" for the x-axis labels. The bin size (0.24 GeV) really looks odd - why not 0.25 or 0.2, which would give an integer number of bins between 0 and 6 GeV?

ecaliso has changed to \cal{I}.

This 0.24 GeV is indeed unfortunate, but it is not possible to change it now, as it would required to redo the analysis.

Also the left "|" before \eta overlaps with \eta, while the right one has a large space - looks bad.

Fixed

Figs. 2-9: all "errors" in the legends should be replaced with "uncertainties".

Done

Figs. 8-9: the "*" index in \theta^* is typeset way too high in the x-axis labels.

Done.

That's all. Good luck with the final review steps and fast publication!

Greg