General comments

Are you planning to increase amount of the analyzed data by including data recorded in 2017 and 2018 ? Yes, there are plans to make analyzis, including 2016, 2017 and 2018 data, dedicated to measurement of pdf of charm quark.

Can your results be combined or compared with those obtained in run 1 ? The results of the measurements depend on the pp collision energy and run 1 and run 2 mearuements were done at different pp collision energy. However, it is possible to compare accuracy of the measurement. E.g. for 8 TeV Z + charm quark measurement (Eur. Phys. J. C 78 (2018) 287) the systematic error for total fiducial cross section is 8% comparing to 6% in current study, and statistical error decreased from 5.8% to 1.3%.

Suggest presenting figures and tables in the order they are mentioned in the text. For example, tables 1-4 are references after Fig. 3, but the order they appear in the paper is very different. Added commands \makeatletter \let\ftype@table\ftype@figure \makeatother to .tex file, which allowed to put tables and figures in the order, which they are mentioned in text. Need to make sure, this command is allowed for submission.

Detailed comments

Abstract: please define eta as pseudorapidity and pT as transverse momentum. Changed "Differential cross sections as a function of the $\pt$..." to "Differential cross sections as a function of the transverse momentum $\pt$..." and "...to have $\abs{\eta}<2.4$..." to "...to have pseudorapidity $\abs{\eta}<2.4$..."

Abstract, third sentence: maybe The final states that contain should be The final states contain ? Changed "The final states that contain" to "The final states contain"

p. 1, section 1, line 4: please define QCD at the first use. Changed "check QCD" to "check quantum chromodynamics (QCD)"

p. 1, section 1, paragraph 2, line 4: suggest using background in singular. In this case background sources. Changed "One of the backgrounds" to "One of the background sources"

p. 1, section 1, paragraph 2, lines 7-8: suggest rephrasing sensitive to the charm quark How can we rephrase this and should we?

p. 1, section 1, paragraph 3, line 3: achieveS ? Fixed, but isn't it better "allows to achieve" ?

p. 1, section 1, paragraph 3, line 5: suggest replacing soft by low-momentum Fixed, is it ok?

p. 2, section 2, paragraph 1, line 7: what is the added information of the text version of the ? Mindlessly copied detector description from other papers, this should be changed, but not sure how.

p. 2, section 3, paragraph 1, line 1: The measurement (singular) of the cross section(S) (singular or plural) is ... Changed to "The measurement of the cross sections..."

p. 2, section 3, paragraph 1, line 1: missing space between fb-1 and pp changed \newcommand{\pp}{{\Pp}\Pp} to \newcommand{\pp}{{\Pp}\Pp\xspace}

p. 2, section 3, paragraph 1, lines 1-2: this is a repetition of what was said at the end of the introduction. Shall we remove it in introduction?

p. 3, section 4, paragraph 2, line 6: impact parameters with respect to what ? added: "...impact parameters with respect to primary vertex.."

p. 3, section 4, paragraph 2, line 7: suggest repeating the sign before 0.05, < 0.05(010) added "<" before 0.5

p. 3, section 4, paragraph 5, line 4: missing space before is Fixed

p. 4, section 4, paragraph 1, line 1: suggest removing GeV after 71, as is done on the last line of page 5. GeV removed after number 71.

p. 4, section 4, paragraph 1, line 4: are quotes needed here ? Other papers (like https://arxiv.org/pdf/1107.4789.pdf) don't use quotes for tag-and-probe, neither should we? Quotes removed.

p. 4, section 4, paragraph 2, line 3: secondary vertices are not yet defined here. This wasn't changed yet: can we change the phrase to "...presence of vertices inside jet, displaced from primary vertix (secondary vertices)..." ?

p. 4, section 5, paragraph 1, line 1: suggest measurement in singular Changed to singular

p. 4, section 5, paragraph 2, line 2: invariant mass of tracks reads like a jargon Is there any better way to say it? Can we change it to "Invariant mass of system of tracks, associated..."

p. 4, section 5, paragraph 2, line 3: IS used (singular) Changed to singular ("is")

p. 5, section 5, paragraph 2, lines 3-4: you have defined SFq, but not SF. You have neither defined SFb, SFc nor SFl. While the first two are quite intuitive, the SFl looks like SF1 and is much less intuitive. Can we define them in the very beginning of the paragraph: "The values of the scale factor SFq for light (SFl), charm (SFc) and bottom (SFb) components, definded as ratio..." ?

Figs. 2, 4: Data/MC histograms are not commented in the caption. Please do it. Can we add to both plots just "Bottom pannels on each plot represent data to MC ratio." ? Do we need to repeat, that shaded area on bottom pannels represents MC uncertainty?

Fig. 2: Is there a significant difference between the data and simulated contributions, e.g. around 1 GeV ? Indeed, there is significant difference between data the and simulated contributions, particulary around 1-2, GeV where contribution from charm component is maximum.

Fig. 2, caption, line 2: missing space after full stop. Space added after full stop.

Tables 1-4, caption, line 2: no need to say in each case Phrase "in each case" removed from these tables caption.

Figs. 3 and 5: what is the purpose of showing accumulated statistics on simulation histograms ? To demonstrate that simulated statistics is the same ? Yes, these errors show that different channels and different components have comparable statistics, and this approach has been used in previous publications, e.g. in the measurement of the differential Drell-Yan cross section, in JHEP 12 (2019) 059

Fig. 3, left, legend: do you mean electron channel ? Indeed, left plot shows distributions of secondary vertex mass for electron channel, muon channel was put in legend by mistake, fixed.

Fig. 3: the distributions of light jets and c jets have quite similar colours, at least on my screen. The light component color in figure 3 was change to more contrast red (kRed+2).

p. 6, section 5, paragraph 2, line 5: missing space between pT and c. There is space between \pt and \cjet in tex, but not in pdf, added ~ before \cjet: ~\cjet.

Tables 1-4: is there expected a smooth behavior of scale factors against pT ? If yes, maybe this fact can be used to reduce the effect of fluctuations and reduce the uncertainties ? Indeed, one could expect smooth dependance of scale factors against different pt bins, however each scale factor is obtained in separate fit of secondary vertex mass which is a separate independant measurement, which could lead to larger statistical fluctiations. In addition, there is non-zero anti-correlation between SFs errors in this fit due to the fact, that secondary vertex mass has already been used for making c-tag discriminator.

p. 7, section 6, item list, first bullet, line 5: would you like to motivate this choice ? Don't know how to accurately substantiate this choice of mu_R/mu_F

p. 8, section 6, item list, jet energy scale and resolution bullet, line 2: secondary vertex mass distributions reads like a jargon Can we use M_{SV} notation, since it has been defined before and shouldn't sound like jargon.

p. 9, section 6, item list, luminosity bullet: missing full stop at the end. Full stop was added in the end.

Table 5, caption: IN or ON ? I don't know if there is a difference in meaning depending on the preposition, of it is just wrong preposition?

Table 5: some table entries miss decimals decimals were added: "1"->"1.0", "3"->"3.0"

Table 5: have the entries in the column Channels been defined ? Can we write the following caption: "Summary of the systematic uncertainties in the integral fiducial cross section arising from the various sources for electron (ee) and muon (\mu\mu) channels, calculated by integrating over pt_{Z} or pt_{c jet}."

p. 10, section 7, paragraph 2, line 1: suggest measurement in singular "Measurements" were changed to singular form "measurement"

p. 10, section 7, paragraph 2, line 3: relatED to "Relating" were changed to "related"

p. 10, section 7, paragraph 2, lines 3 and 6: suggest replacing slash by AND Slash changed to "and" in these lines.

p. 10, section 7, paragraph 3, line 3: missing comma before respectively Comma added before respectively, after uncertainties .

p. 11, section 8, paragraph 1, line 5: background sources "the main backgrounds" changed to "the main background sources"

Tables 6 and 7, caption, lines 1-2: combine channels does not read well. Maybe electron and muon channels and combined value. or electron and muon channels and channel combination. Changed "...for electron, muon and combine T channels..." to "...for electron and muon channels and combined value..." in caption for tables 6 and 7

p. 12, section 8, paragraph 1, lines 5: Suggest removing SINCE "Since" removed in this line.

References [27]: unnaturally long dash in the title. There is probably a misunderstanding, because reference [27] doesn't contain dashes, however, reference [28] does contain long dash, which had been changed. We use " title = "{\GEANTfour}---a simulation toolkit", in .bib file, changed "---" to "--", that is how inspirehep.net suggest to cite. "

Appendix A Figs. A1, A2: please define bottom plots in the caption. Can we add "Bottom pannels on each plot represent data to MC ratio." just like for figures 2 and 4 ? Figs. A1, A2, caption, line 1: secondary vertex mass not defined. Changed "secondary vertex mass" to "M_{SV}" which is used in text. Can't find the rule: do we need to define it again in appendix? Figs. A1, A2, caption, line 2: missing space after the full stop Added space after full stop.

abstract, line8: "...at least one must have pt > 30..." Thats inaccurate definition of feducial volume: leading leptons must have pt > 26 GeV, fixed.

l27: "...as a function of the c-tagged jet pt..." Differential cross-section is a function of c jet pt, not c-tagged jet pt, fixed.

table5: outdated values, bugs in calculating MC errors, fixed (see Table5Fixed.pdf).

In general

(1) All Figures (except Fig.1), in the legends, axis and plot labels, (a) the 1st letter of each line should be in the upper case; (b) the non-leading words and letters should be in the lower case, e.g.

(i) Fig.2(left) and Fig.3(upper)

** the legends: (the 1st lines)

"Data (Electron channel)" --> "Data (electron channel)"

(ii) Figs.4-5

** the legends:

"muons electrons" -->

"Muons Electrons"

(iii) The other places where the similar changes should be made are

Other plots of Figs.2-3's legends, Fig.4-5's vertical axis labels, and Figs.7-8's legends.

(2) Throughout the paper (including in the Tables, etc.), to be consistent with the "c-tagged ..." everywhere in the paper, also per the PubComm guidelines, the hyphen is allowed in between "c tagging" if it plays a role of adjective, e.g.

L133-134: "to measure the c tagging efficiency for c jets," --> "to measure the c-tagging efficiency for c jets,"

Other places where the hyphens should be added in between "c tagging" are

L137, L140-141, L159, L197, L210 (two places), L213, Table 5 (the header row and the 4th column), L248, and L251, etc.

Pages 1-4

(3) L7, as the "IC" has been used for only one time at the end of L7 in whole paper, it may not be necessary to be introduced at the 1st half of L7, then it can can be simply spelled out at the end of L7, i.e.

"of observing the intrinsic charm quark (IC) component in the nucleon [1â€]. An IC" -->

"of observing the intrinsic charm quark component in the nucleon [1â€]. An intrinsic charm quark" or "of observing the intrinsic charm quark component in the nucleon [1â€]. Such an intrinsic"

(4) L8-9, L11-13, L64, L89, L93-95, L100, L106, L127 and L142. These lines may be shortened from

(a) L8-9: (as the "pT" has been introduced on L5)

"at large values of the transverse momentum of the Z boson and of the c jet." -->

"at large values of pT of the Z boson and of the c jet." or "at large pT of the Z boson and of the c jet."

(b) L11-13: (as the "LSP" and "SM" have not been used afterward in whole paper) L11: "for physics beyond the Standard Model (SM)." --> "for physics beyond the Standard Model."

L12-13: "an undetected lightest supersymmetric particle (LSP)," --> "an undetected lightest supersymmetric particle,"

(c) L64: (two places; the 1st place is to follow the good examples on L55 and L117, etc.; the 2nd place is to use the "NLO" introduced on L57)

"Z + jets cross section calculated at next-to-next-to-leading order with ..." -->

"Z+jets cross section calculated at next-to-NLO with ..."

(d) L89: (to be consistent with all other CMS papers, it will be looked better if the superscript of the power "2" of "pT 2" is moved toward left to be above the subscript "T")

"pT 2" --> "p2T"

(e) L93-95: (two places; the 1st place is to follow the good examples on L110, etc.)

"ranges from 1.7% to 4.5%. The dielectron mass resolution for Z -> ee decays when both electrons are in the ECAL barrel is 1.9%, degrading to 2.9% when both electrons are in the endcaps." -->

"ranges from 1.7 to 4.5%. The dielectron mass resolution for Z -> ee decays when both electrons are in the ECAL barrel is 1.9%, degrading to 2.9% when they are in the endcaps." or "ranges from 1.7 to 4.5%. The dielectron mass resolution for Z -> ee decays when both electrons are in the ECAL barrel is 1.9%, degrading to 2.9% when in the endcaps."

(f) L100 and L106: (to follow the good examples on L99, etc.)

L100: "efficiency to reconstruct and identify muons is greater than 96% [31]." --> "efficiency to reconstruct and identify muons is > 96% [31]."

L106: "to be less than 25% of the muon pT." --> "to be < 25% of the muon pT."

(g) L127: (two places; the 1st place is due that the "mZ" has been introduced on L72-73)

"close to the mass of the Z boson: 71 GeV < mee/mumu < 111 GeV." --> "close to the mZ: 71 < mee/mumu < 111 GeV."

(h) L142: (two places; the 1st place is similar as the 1st part of the item (e) above)

"between 5% and 15% depending on the pT of the jet." --> "between 5 and 15% depending on the jet pT."

(5) L60, the "QCD" should be explained at its 1st appearance in text here, i.e. "calculated at NLO in perturbative QCD." --> "calculated at NLO in perturbative quantum chromodynamics (QCD)."

(6) L103-104, to follow the good examples in many CMS published papers,

e.g. in SMP-20-001 (Page 5), SMP-19-014 (Pages 2-3), SMP-19-012 (Pages 4 and 8), SMP-19-008 (Page 3), SMP-19-001 (Page 4), SMP-19-014 (Pages 2-3) and SMP-17-012 (Pages 5-7), etc.

the "as the sum of the pT of all ..." can be shortened to "as the pT sum of all ..."

(7) L169, to be consistent with all other CMS papers, the "Section xx" in text should not be abbreviated,

"are discussed in Sec. 6." --> "are discussed in Section 6."

Page 9, Table 5

(8) In the header row and the 8th column, the 2nd word should be in the lower case, i.e.

"Top Pair" --> "Top pair"

Page 10, Fig.6

(9) In the legends and the axis labels, the letter size seems too small, almost invisible. Readers must appreciate it if the letter size can be enlarged as much as possible.

(10) In the vertical axis labels,

(a) to be consistent with the horizontal axes of this Figure and all other Figures in this paper, the unit should be put into the square brackets instead of the round ones, i.e.

"(pb / GeV)" --> "[pb / GeV]"

(b) to be consistent with Figs.2-3 and 7-8, etc., I'm not sure whether the lower panels should be swapped from

"MC/Data" --> "Data/MC"

If so, the data points and the bands in the lower panels should be flipped vertically too. In this case there are many different MC models, so MC/data value is plotted.

Page 11, in the Summary Section

(11) L272-274, L276 and L278-281. The "NLO", "QCD", "PDF" and "MC" should be explained (or spelled out if it would not be used again in the Section) at their 1st appearances in the Summary Section on these lines, since some readers may only read the Summary Section instead of whole paper, i.e.

(a) L272-274: (four places)

"at next-to-leading order predicts 524.9 +- 11.7(th) pb for the same fiducial region. The theoretical systematic uncertainty includes QCD scale and PDF uncertainties. The predictions from several MC generators ..." -->

"at next-to-leading order (NLO) predicts 524.9 +- 11.7(th) pb for the same fiducial region. The theoretical systematic uncertainty includes quantum chromodynamics scale and parton distribution function (PDF) uncertainties. The predictions from several Monte Carlo generators ..."

(b) L276 and L278-281: (five places can be shortened)

"SHERPA at next-to-leading order tend to overestimate the cross section ... with FEWZ at next-to-next-to-leading order. Since the prediction of ... at next-to-leading order is in better agreement with data than that ... could be an indication that the parton distribution functions ... ... existing constraints on the parton distribution function of the" -->

"SHERPA at NLO tend to overestimate the cross section ... with FEWZ at next-to-NLO. Since the prediction of ... at NLO is in better agreement with data than that ... could be an indication that the PDFs ... ... existing constraints on the PDF of the""

Pages 12-15, in the References Section

(12) L330-331, in [1], to be consistent in this Section, the journal name should be shortened from

"Advances in High Energy Physics 2015 (2015) 231547," --> "Advan. High Ener. Phys. 2015 (2015) 231547," or "Advances in HEP 2015 (2015) 231547," Fixed: bibitems were taken from inspirehep

Other ones which also need to be changed by the similar way are

[2] (shortened to "JHEP"), [3] and [46] (shortened to "Euro. Phys. J."), and [38] (shortened to "JINST").

(13) L334, in [2], to be consistent in this Section, (a) an extra index after the year number should be removed, (b) the volume index should be corrected, i.e. (together with the item (12) above the for journal name)

"Journal of High Energy Physics 2018 (2018), no. 2, 59," --> "JHEP 02 (2018) 059," Fixed

Other ones which also need to be changed by the similar way as (a) are [3], [38] and [46]. Fixed: bibitems were taken from inspirehep

(14) L343, in [5], to be consistent in this Section and this paper, an extra space before the symbol "+" in the article title should be removed, i.e. "Z +charm production in proton-proton" --> "Z+charm production in proton-proton" Fixed: bibitems were taken from inspirehep

(15) L433-434, in [38], to be consistent in this Section, all references should have only one page index instead of two, i.e. (together with the items (12)-(13) for the journal name and the extra index) "Journal of Instrumentation 15 (2020), no. 09, P09018–P09018," --> "JINST 15 (2020) P09018," changed, bibitem copied from inspirehep

(16) L439, in [40], to be consistent in this Section and this paper, the particle symbols in the article title should use the font of non-italic, i.e. "Measurement of BBbar(italic) angular correlations based on ..." --> "Measurement of BBbar(non-italic) angular correlations based on ..." changed, bibitem copied from inspirehep

(17) L451, in [44], to be consistent in this Section (e.g. [23], etc.), the document name should be changed, and the names of institute and city should be removed before the year number, i.e. changed, bibitem copied from inspirehep

"Technical Report CMS-PAS-LUM-17-001, CERN, Geneva, 2017." --> "CMS Physics Analysis Summary CMS-PAS-LUM-17-001, 2017." changed, bibitem copied from inspirehep

Page 16, Figs.7-8

(18) The captions of two Figures are almost identical except differing by only one word at the beginnings of the 2nd line; thus two Figures may be combined with an extended 2nd line of Fig.7's caption, i.e. (and a space should be added after the 1st sentence)

"Figure 7: Distribution of the secondary vertex mass of the ..., for electron channel.The observed ..." -->

"Figure 7: Distribution of the secondary vertex mass of the ..., for electron (upper row) and muon (lower row) channels. The observed ..."

On the other hand, if no combination would be considered, at least one change should be made, i.e.

a space should be added on the 2nd line of each caption)

"channel.The observed data is compared to ..." --> "channel. The observed data is compared to ..."

(19) In each legend of Fig.7, the 2nd word on the 1st line should be changed from (together with the item (1b(iii)) above for the lower case)

"Data (Muon channel)" --> "Data (electon channel)"

1. Do the JER and JES uncertainties quoted apply to c-jets as well as inclusive jets? (e.g. the uncertainties quoted on lines 219 to 221

We apply JER and JES uncertainties to all jets: this may have impact not only on shape of the pt spectra of selcted c-jet , but on the content of the event, especially for events with c-jet pt close to the cut threshold.

2. Figure 1 a. What is the brown color in the stack? I cannot find where it is defined. b. It would be easier to follow if the order of legend matched the order of the stack c. It is important to see the shape differences between the c-jet and the other contributions. If I understand correctly this is the basis of the fit. The stacked format makes it very hard to compare the shapes. Could you include the non-stacked shapes, each normalized to unity as an inset or as a separate figure?

Fixed

3. Results: If I understand correctly, you only show results in the fiducial region, and never extrapolate to the full phase space, correct? The intro to section 7 could be misleading where you say that dsigma/dpt is the differential cross section ahd sig_fid is the fiducial cross section. I would remedy this by adding "fiducial" to the first sentence (line 240). We defined the fiducial cross-section in the same section 7, but perhaps it is worth discussing at FR

4. In Fig. 6 the bands on the theoretical predictions are never explained. From the legends, I am guessing that these represent renorm/fact scale variations and PDF variations. This can be confusing because you already say that scale and PDF uncertainties are part of the experimental systematic uncertainty. My guess is that the experimental uncertainties include the effect of PDF and scale on the acceptance, while the bands on the theory include the effect of these same things on the cross section itself. I think this is a common way of presenting this, but it needs to be explained explicitly. Also, you should explain why you do not include such uncertainties on the LO prediction.

It was added to the table, that NLO predictions have PDF (for Sherpa) and PDF + QCD (Madgraph) uncertainties. There is also a mistake in the draft: QCD and PDF uncertanties are theoretical, not experimental uncertainties.

In the reactions c quarks and not c jets are produced together with Z, the jets are observed in the events. This should be corrected everywhere in the text: in lines 4, 8, 10, 13, 14, 16, 26, 57, 262, and in the caption of Fig. 1. Thus "c jet events" (I should prefer to write it as "c-jet events" as it is a modifier here) are OK, but "Z + c jet production" is not. At different stages of the ractions different physics objects are produced: if we measure c-jets, then in my opinion it is better to write about jets, since c jets can be also produced from gluon splitting.

Abstract: The sentence "Differential cross sections with respect..." is awkward and it should also emphasize our priority in this measurement.. What about: "Differential cross sections are measured vs. the transverse momentum of the Z boson candidate and that of the charm jet at the first time, and compared ..." The abstract had been changed, but still worth discussing at FR

L53: "with 23 additional interactions (pileup) " => "with 23 interactions (pileup) ", we did not define the primary vertex yet, so cannot say additional. It is OK in L77. Fixed

L78: "interactions (pileup) in" => "interactions in" pileup was defined in L53 We removed definition at l53 and added to L78

L81: "particle candidates" => "particle candidates (physics objects)" to define it Fixed

L94: Naive question: How could the dielectron mass resolution for Z → ee decays in the endcaps be so much better than the momentum resolution for electrons with pT ≈ 45 GeV? Need to think / check more papers. Copied from another paper without thinking through

L99: "<0.5 and" => "less than 0.5 and" Fixed

LL120-122: All that jet removal sounds a bit voluntary, it may introduce syst uncertainties, as the criteria are not really defined. This is a standard sentense for object reconstruction section

L126: "a pair of opposite-sign electrons or muons satisfying identification and isolation criteria, with pT > 10 GeV and |η | < 2.4," This must be for the leptons, not the vectorial sums: "satisfying" => "each satisfying" Fixed

L167: "scale factor SF_q": add "(q stands for uds=\ell, c, or b)" . Do we need additional explaining of 'q' ? We write later, that these are for Z+q components, which had been already defined: Z+light, Z+c and Z+b

Fig. 2: There is a quite significant deviation at around 1 GeV, as if something were missing. The scale labels of the axes are too close to the borders of the plots. Plots have been redone. The deviation is due to excess of c jets events in MC, this leads to the disagreement between data and MC, described in results and summary.

L208: "different PDF replicas": Is "replica" an official term, not jargon? It is used in other papers as well.

Table 5: How were these syst. uncertainties summarized? Are they statistically independent or their correlations were taken into account? We describe further in the text how these uncertaitnies are used in combining channels, and which of them are considered correlated between channels.

Funny to see in L237 the caption of Table 5 repeated. Why do not we say instead in the text: "The uncertainties on the inclusive fiducial cross section from the considered sources are listed in Table5." Fixed

L244: " integral fiducial purity and ε is the integral fiducial selection efficiency" What does integral mean here? Average over the fiducial region? One cannot integrate purity or efficiency... Integral means for integral cross-section, calculated from the integral number of events. I thought, integral efficiency is a used term.

LL262-264: I propose for the first sentence "Differential cross sections for inclusive Z+c jet production as a function of the transverse momentum pT of the Z boson and of the c jet have been measured for the first time using 35.9 fb−1 of data collected by the CMS experiment in proton-proton collisions at \sqrt{s} = 13 TeV of the LHC." The sentence had been changed, but maybe it is worth discussing

L266: "(at least one with p T > 26 GeV)," This does not agree with L128 where the electron was supposed to have pT > 29 GeV, please specify correctly, 26 for muons, 29 for electrons. There is no contradiction here: the signal is defined at gen level, where electrons and muons have the same cuts. On reco level however, the selections are different for leptons, in particular pt for leading electron is > 29 GeV

L274: "with measurements" => "with the present measurements" I agree, that measurements should be specified, but what is the better wording: "present", "these" or just "the" ?

L275: "shows good agreement with data" => "shows a good agreement with the measurements" ("data" is jargon here) The whole sentence had been changed

L279: "with data" => "with the measurements" Data seeems to be a standard term, maybe worth to be discussed.

The comments are in SMP-19-011-v15-FR-Nov13-2.pdf file, attached to the page. Almost all suggested corrections are done in the text.

Line 7: I am not sure what “intrinsic” means. Does such a concept make sense? Intrinsic component (e.g. intrinsic charm component) is a standard terminology.

Line 14-15: I find this sentence oddly formulated. Would not a sentence such as: “Studying the visible decay modes of Z+C jet processes will improve the modelling of such backgrounds.” Be clearer ? The sentense was changed

Line 53: “…ns with 24 interactions on average per beam crossing”, leaving the definition of pileup to line 78. Fixed

Line 89: I’m am not sure if “physics object” is acceptable jargon, but I think it could probably be defined in the first paragraph of Section 4, as outputs from the PF algorithm? It is a very common sentence, used in other papers.

Lines 93-94: This sentence implies these mass resolutions are constant over EE and EB.

At least say “about” or “around”, or better, find a formulation giving ranges or perhaps

maximum values. It is a very common sentence, used in other papers. Do we need to add "around/about" ?

In Figs 2 & 3, modulo how the plots appear on my screen and paper version, the

grey areas are not identified. I may guess they represent MC uncertainty, but in

the upper plots they are split into two shades of grey which is puzzling.

Also there is no statement of what the data error bars represent. It is not grey area, but dashed area, it is a common way to show MC uncertainties, also used in other papers.

Fig 4, left. Presumably three of the muon data points are hidden

behind the electron ones. Cannot the plot be re-jigged to show this?

Eg in the old days, the error bars would be slightly displaced. Fixed

Table 5: Given the previous bullets describing the estimation of the

errors I expect that the last column should be labelled M-sv. Also the

caption is slightly inaccurate if this is a statistical error. It is always a question, where MC stat error should be assinged. We desided, that since MC statistics have impact on results of the fit of SVM, we put mc stat errors as a different source of uncertainty.

Fig 6 is not easy to understand. In the first place it would benefit from

being significantly larger. Second, in the right plot, the width of the third

bin appears to be truncated by the inset. And the meaning of the error

bars on the data points should be given. This plot (now #7) was redone.

Trivia:

Cover page: Don’t know if we comment on this but under PDFKeywords is written “your topics”. Which keywords do we need here?

Line 18: “A previous….” Fixed

Line 87: I would put a comma after “trajectory” do we need comma here?

figures 2-3: please make larger (and maybe bolder?) the axis labels, axis digits, the legend, the "CMS" (top left), the data points (in plots and the legend); please move x-axis labels away from the axis; please add in the caption what is represented by the data horizontal bars. Fixed

figures 4-5: please make larger (and maybe bolder?) the axis labels, axis digits, the "CMS Simulation" (top left), the data points (in plots and the legend). Fixed

figure 6: i suggest to have this figure as 2 vertically placed panels and make it as big as possible (the whole page) - this is a main result of the analysis and the figure now is poorly visible; Fixed

table 5: please align the numbers - like in other tables. please make larger (and maybe bolder?) the axis labels, axis digits, the legend, the "CMS" (top left), for the legend - you may try to break it into 2 lines for some curves, if necessary; the data points (in plots and the legend); please move x-axis labels away from the axis; why there is no band on the clue curve? what exactly are the bands on green and red curves? i suggest to remove the grid from the ratio plots and make the gray band darker (it is hardly visible now, especially in the legend). the figure is very busy, i suggest to remove the yellow band - representing the data uncertainties - and add horizontal bars to the data points (representing the same that the yellow band); please add in the caption what is represented by the data bands (or horizontal bars) and the theory bands; Table was changed

line 78: please add a standard sentence that the MC events were reconstructed/treated in the same way (with the same programs) as the data events. "Fixed: The simulated events are reconstructed with the same algorithms as used for the data."

line 203: why is this uncertainty labeled as "QCD renormalization scale" if it concern both, the renormalization and factorization scales? please unify. Fixed

line 207: please state if the largest uncertainties or a sum of uncertainties from different scales is taken as this systematic uncertainty. "Largest deviations from the central values were set as uncertainty."

line 208: please state which additional PDF sets are taken for this systematic uncertainty. DY NLO sample includes only one PDF set, but since the value for each x parameter of the pdf is known with some error, there are multiple replicas of the same PDF, corresponding to variation within this error.

section "results" in general: please describe what is meant by theory uncertainties for the predictions mentioned in line 259, as well as for the theory predictions bands in figure 6; if there is no uncertainty for the blue prediction, please state why; if the uncertainty is very small, please state that as well. we don't use uncertainties for madgraph LO

lines 256-260: move to line 253, before the description of figure 6. Need to be discussed: the idea was that we first discuss differential cross section, then integral cross section.

line 258: ... uncertainties, respectively. is coma really required here?

line 259: i think you mean "The tehoretical systematic uncertainty includes ..."? otherwise it makes no sense. yes, Fixed

line 255: please add a description of how the various theoretical predictions describe the measured cross sections. right now there is no comments about it at all (for the differential cross sections). this is stated in lines 274-280 of the summary - should be first mentioned here (possibly in a more elaborate way). added: The prediction from \mgamc at leading order shows good agreement with data, while both \mgamc and \SHERPA at next-to-leading order tend to overestimate the cross section.

Abstract

lines 7-10: The details of the selection process do not need to be in the abstract. Yes, selections are discussed in other sections. But we decided to add fiducial volume definition to abstract to give an idea, for which kinematic parameters the signal is defined .

Text

line 99: Why would muons and electrons have a different impact parameter? We use standard muon and egamma pog suggestions for selections of leptons, which may differ for electrons and muons. Possible explainantion is that impact parameter cut is applied along with other criteria, on the other hand it must fit predifined efficiency of the selections.

line 151: Do you really add photons to the muon tracks? Maybe change "leptons to "electrons". We apply the same algorithm to generator muons and electrons .

line 199: How are the input parameters varied? It depends on the source of the uncertainty, some details are given for each separate source. In some cases these are just +/- error for some paramters, for PDF and QCD these are sets of weights, corresponding to different mu_R and mu_F and different PDF replica

line 268,269: How can unfolding be used to take into account detector effects? Maybe you mean convert observable measurements into the signal phase space. I agree, that is more convenient to think about unfolding as a reverse transformation from one space (detecter level measurements) to another (generator level). However the phrase about detector/resolution effects is common in context of unfolding, so maybe it is better to keep it his way

Abstract

It is a bit unusual to give the cuts (lines 7-10) in the abstract. On the other hand I understand you need them to define the fiducial region for the inclusive cross section result.

I suggest to push these lines to the second part of the abstract which would read as follows

Measurements of differential cross sections of the production of a Z boson in association with at least one jet initiated by a charm quark in proton-proton collisions at sqrts = 13 TeV are presented. The data were recorded in 2016 by the CMS experiment at the LHC and correspond to an integrated luminosity of 35.9 fbâ^'1. The measurements are performed in final states that include a pair of electrons or muons consistent with being the decay products of a Z boson, plus a jet consistent with being initiated by a charm quark produced in the hard interaction.

Differential cross sections with respect to the transverse momentum of the Z boson candidate and with respect to the transverse momentum of the charm jet are measured, and compared with predictions from different Monte Carlo generators.

The inclusive production cross section is measured in a fiducial region requiring both leptons to have |η| < 2.4 and pT > 10 GeV, at least one lepton with pT > 30 GeV, and their invariant mass in the range 71â?"111 GeV, while the c jet candidate is required to have pT > 30 GeV and |η| < 2.4. It is measured to be 405.4 ± 5.6(stat) ± 23.5(exp) ± 8.1(th) pb.

These are the first measurements of this process in proton- proton collisions at 13 TeV.

Fixed.

L 18: It would be interesting to tell the difference with the approach of [5]. It is not only the energy and lumi which are different (this is not so crucial), but the use of an inclusive approach to c jets rather than selection exclusive channels like D or D*. This allows to achieve bigger selection efficiency comparing to $8\TeV$ study, where c jets were identified by reconstructed $D^{*}(2010)$ meson or soft muon inside the jet.

Lines 26-34: in my opinion, these should not be in the introduction. What about pushing these in section 4 (for example after L143 or line L152. Furthermore, L 26-31 are pretty trivial, telling us that you have to select events and bin them in PT. Isnâ?Tt it obvious, to measure differential cross sections?! The only useful information is about the unfolding and that you combine electrons and muons results. This is the strategy of the measurements, without details. I think it should be there for those, who will read only introduction to have and idea, what was done.

L103-106: why is there a text on muon isolation, but nothing for electron isolation in the previous paragraph? Are you not asking for electron isolation? The formulation lines 127-128 is ambiguous, whether isolation applies only to muons or to both. Fixed. Isolation for both muons and electrons discussed in separate paragraph (as it is done in other papers)

L159-161 â?osufficient differencesâ?� . THIS IS THE KEY OF THE ANALYSIS. How can one believe you on the basis of Figure 2? I think it is IMPERATIVE to show the normalised distributions of the 3 species (light-jet, b jets, b jets) to convince the reader that you can separate them. I actually measured each coloured bin of Figure 2 and plotted them (on a linear and on a log scale !) to convince me that there was indeed some discrimination left. For fun, I attach the plots below . The plots were added to the paper, along with correlation factors. The largest anticorrelation is for charm and bottom components, which is ~-0.85.

Line 167etc.. Scale factors

I was initially confused by the wording. SFs refer usually to corrections of efficiencies or resolutions, while for physics results I am talking more of â?oweightsâ?�. Let us see if other members of the pubcom had the same problem.

i am wondering whether all these tables 1-4 are useful. First, a theorist cannot use SFc without knowing the cross section in a particular bin predicted by the Monte Carlo generator. Much more useful would be a table of the values plotted in Figure 6. Furthermore SFl and SFb are useless information for the reader, except to say that they are compatible with 1. Fit of SVM is one of the key parts of the analysis, hense these fit results should be added to the paper. Table for fig.6 (7 in new version) was added to the suplementary material.

By the way, how correlated are these 3 factors? Also surprising are some up-and-downs of SFl and SFb (some times as large as 30%) from one bin to the next, 3 to 4 statistical sigmas apart; and it is difficult to judge whether the systematics (certainly correlated between successive Pt bins) can explain these jumps. The correlation varies from ~ -0.85 for errors SFb and SFc to -0.25 for SFc and SFl

Table 5

Putting the positive value and the negative value on different lines is not very aesthetic. In addition, the two values are most of the time very close. Who cares about the small difference between them ? Why not giving simply an average (unsigned) estimate? Anyway, this is only to give an idea of the various contributions. I agree, that table with one value - average, instead of +/- looks cleaner, changed

L 256-260. Why are these lines (which refer to the inclusive X-section) not after L245? Fixed.

Summary: the discussion of lines 273-280 should be better placed in the section 7 which could be called â?oresults and discussionâ?�. It is unusual to have such discussion (including a reference, [46] ! ) in the summary. Maybe worth discussing at FR

Figure 6 IS THE RESULT OF THE PAPER ! Was it not possible to make it more legible, possibly by separating the 2 parts?

Plots were updated, ~2 times higher

Concerning the updates (slides attached):

• the changes to the statistical error for e/mu channel separately (slides 2-3) looks ok to me (except that bottom left plot in slides 3 I think is the one obtained with RooFit and not with TFractionFitter). In the new plots the central value is still the one obtained including MC errors right?
• on slide 4 if I compare the combined plots with those obtained with RooFit, I still find that the statistical error is much larger, especially in the c-jet pt distribution. Do you understand why? The statistical error still refers only to statistical error in the data? Combine it is now run with systematic errors fully correlated? (as it is written in lines 244-246 of the new draft)

Z pt stat errors for latest version are ~equal to stat errors in version, obtained with RooFit. For c-jet electron channel stat errors are bigger, since no regularization used for unfolding.

About the text here are a few corrections (english wording on the suggestions to be reviewed by Joel)

• line 5-6: Looking at CMS-SMP-15-009 this part on the intrinsic charm can be expanded: split the present sentence after nucleon and change it to â?oand possibly observing the intrinsic charm quark (IC) component in the nucleon. An IC component would enhance the Z+c production, in particular at large values of the transverse momentum of the Z boson and of the c jet." then add these references after "nucleon":

https://arxiv.org/abs/1504.06287 <https://arxiv.org/abs/1504.06287> https://arxiv.org/abs/1605.06515 <https://arxiv.org/abs/1605.06515> https://arxiv.org/abs/1707.00657 <https://arxiv.org/abs/1707.00657> Then we can start a new paragraph with â?oAssociated production of a Z boson and a c jet is also an important background in searches for physics beyond the Standard Model (SM). For example, in supersymmetry â?¦" fixed

• line 9: do we really need reference [1] for SUSY models?

removed

• line 10-11: in the sentence â?oOne of the main backgrounds for such a process is Z+c jet production with missing transverse energyâ? change â?owith missing transverse energyâ? to â?owith the Z decaying into neutrinosâ?.

fixed

• line 12-13: change to â?owill therefore enhance this and similar searchesâ? in order to be less specific

fixed

• section 3: avoid new paragraph after line 55 and 60 and split sentence at line 61 after "PDF setâ?. The rest of the paragraph, starting with â?oThe program GEANT4â?¦â?, should be moved at the end of the section.

fixed

• line 95, 175 and 176: â?o<â?o and â?o>â? I think should not be used without a variable on the left, i.e. instead of â?olower(greater) thanâ?.

Changed at 95 to "greater than", bot for 175 and 176 not sure, if it can be said this way. changed to "...with \PQc jets with generated \pt ${<} 30\gev$ but reconstructed \pt ${>} 30\gev$ ..."

• line 107: â?oto correct forâ? -> â?oto removeâ? (to avoid repetition of â?ocorrectionâ? and â?ocorrectâ?)

fixed

• line 110: "particle level jetsâ? -> â?ogenerator-level jetsâ?

fixed

• line 131: in â?opresence of secondary vertices and tracksâ? there is no attribute to the tracks: maybe â?odisplaced tracksâ? or â?otracks with large impact parameterâ?

fixed, changed to "tracks with large impact parameter"

• section 5, first paragraph can be improved in my opinion. I suggest to changi it from line 151 to â?oâ?¦ in the c-tagged jet [5] can be used to discriminate between signal and background. Figure 2 shows the observed distributions of M_SV in the electron and muon channels, compared to the different signal and background contributions predicted by the simulation. Even if the discrimination power of this variable is reduced for c-tagged jets, each flavor component can be measured by fitting the distribution of M_SV.

fixed

• line 165 â?ouncertaintiyâ? -> â?ouncertaintyâ? fixed

• line 167: â?oobserved" -> "observed" fixed

• line 168: â?oand dataâ? (remove â?oobservedâ?) fixed

• line 268-269: change to "while both Madgaph5_aMC@NLO and Sherpa at next-to-leading orderâ?

fixed

• One last thought: maybe we can add at line 271, just before the last sentence â?oSince inclusive production of Z boson associated with jets at next-to-leading order is better agreement with data than at leading order [https://arxiv.org/abs/1804.05252], this can be an indication that the charm pdf overestimates the charm content."

fixed

• Changed from RooFit to TFractionFitter for fitting SVM distributions.

• If two SV found inside jet, one with larger significance is chosen.

• Luminosity uncertainty taken into account at ee and mumu combining step.

Plots with short summary of updates are in updates.pdf

Greg Landsberg

Abstract, L6: of a Z boson, and a jet consistent;

Introduction:

L2: The CERN LHC has delivered large samples; L6: parton distribution function (PDF); L9: add a comma before "leaving"; L13: to the charm quark PDF, is; LL14-15: of the Z+c jet production cross sections at s√=7s=7 and 8 TeV can be found in Refs. [3] and [4], respectively. L16: pair of electrons or muons; LL16-17: to have originated by a charm quark by applying charm tagging criteria [5] (c-tagged jet). LL18-19: to the generator level. L21: add a comma before "as a function"; L22: add a comma before "are"; L23: a c jet enriched [or a c-jet-enriched] sample;

Data and simulated samples:

L47: Drell--Yan [en-dash, not a hyphen]; L50: (light-flavor jet). LL53,56: pp →Z+n jet [make the typesetting consistent with that on L50]; L55: In addition to theMG5\_aMC samples, the \SHERPA [11,12] event generator; L58: In addition to events with light-flavor and b jets, there is; L59: These samples were generated; LL68-69: the NNDPF3.1 [26] PDF set, and \GEANTfour [27] is used for the CMS detector response simulation. LL69-70: pp interaction in the same and nearby bunch crossings (pileup).

Object reconstruction and event selection:

L72: The particle-flow (PF); L74: add a comma before "while"; L85: the pT of all additional PF candidates; L98: \FASTJET [use the pen name!] package [31,32]. L104: nonlinear and nonuniform [CMS Style]; L105: particle-level jets; L122: measured using the ``tag-and-probe" method [37]; L126: and light-flavor jets; L128: and mistag rates for the b and light-flavors jets. L132: 1.2\% for light-flavor and 20\% for b jets. LL136-137: of the hadrons within the jet. L140: as Z+light-flavor jet (Z+LF) events.

Signal determination and unfolding:

L143: Signal extraction and data unfolding; LL149-150: for the Z+b and Z+LF jet backgrounds; L152: subscript "q" in Roman; L153: Z+q jet [q in Roman]; LL153-154: Sources of systematic uncertainties; L156: and the pT ranges in which; Tables 1-4 captions, L1: Values of the Z+LF, Z+c, and Z+b jet scale factors; L2: as functions of; L3: from the fit, while the second one is the systematic uncertainty. Also, move all the captions above the tables [CMS Style]. Tables 1-4 bodies, header line: SFLF [subscripts in Roman]; remove all the vertical dividers [CMS Style]; Fig. 2 caption, LL1-2: of the secondary-vertex mass (mSV) of the highest pT [superlative compound modifiers are not hyphenated!] c-tagged jet, for the electron (left) and muon (right) channels. L164: because of the finite detector resolution. L168: for the electron and muon channels. L172: \textsc{TUnfold}; L174: add a comma before "as a function"; L175: Z boson or c jet pT for the electron and muon channels.

Systematic uncertainties:

L178: Systematic uncertainties are; L184: μRand μF[subscripts in Roman]; L189: in the c-tagged jet rats is estimated; L191: mistagging b and light-flavor jets; L195: the secondary-vertex mass; SFb and SFLF measurements. L197: The JER uncertainty [can't start a sentence with an acronym!]; L198: coefficient in simulation up and down; L202: the identification (ID) and isolation (Iso) of electrons and muons are; L205: than 2 and 1\%, respectively. L206: Top quark pair production cross section; L207: of top quark pair production; L209: Integrated luminosity: The uncertainty is obtained by varying the integrated luminosity; Fig. 4 caption, L2: for the electron and muon channels, as functions of; Fig. 5 caption, LL1-2: for the electron and muon channels, as functions of; L2: and c jet (right).

Results:

L213: unfolded distributions as:; Eq. (1): end the equation with a comma; the numbering is not needed, so suggest dropping "(1)"; Table 5 caption, L1: uncertainties in the; move the caption above the table. Tabl 5 body, header line: QCD -> Scales; Top Pair -> Top quark; swap the electron and muon rows in the table; in the electron rows, "ee" in Roman in the first column. Remove all the vertical dividers [CMS Style]; L216: with ℓ= e or μ L217: The results from the electron and muon channels are; \textsc{Convino} tool; L219: pileup, and top quark pair production cross section sources. L220: and c jet, after the combination, are shown in Fig. 6. LL222-223: (LO), and \SHERPA. The inclusive fiducial cross section value for the Z boson pT<300vGeV; LL223-224: add spaces before every opening parenthesis and use "(theo)" [CMS Style];

Summary:

LL228: using data collected by the CMS experiment at the LHC, corresponding to an integrated luminosity of 35.9 fb−1vLL230-231: for the leading lepton, pT>10 GeV for the subleading lepton, |η|for both leptons, and; L232: were Z+light-flavor jet, Z+b jet, top quark pair, and; LL233-234: for the signal with the Z boson pT<300 GeV; LL234-235: add spaces before every opening parenthesis and use "(theo)" [CMS Style]; L239: to extract the parton distribution function of the charm quark;

Polikarpov:

L14: remove second "Sqrt(s)="

L15: "Here" -> "In this paper"

L19: the generator

L19: suggest to reword as "The analysis uses the data set corresponding to ..... , recorded by ..... 2016."

L27: Finally, the unfolded distributions in muon and electron channel are combined and compared with ...

L55: [11][12] -> [11,12]

L123: line ended with comma.

L125: tagging criteria as provided by the CSV algorithm [5].

L131: .. analysis corresponds to the tagging efficiency of ...

L152: "... scale factors SF_q, " may be confused with SF_{q'}, i.e. a scale factor of q' quark. Consider rewriting to avoid confusion: "The values of the scale factors SF_q are defined as the ratio ...... Z+q jet process. The measured values are given in Tables 1-4 and the sources of systematic .... "

L156: remove extra space before dot

L163-164: these lines do not read well, consider rewording as "primarily due to c jets with generated pt below 30 GeV having the reconstructed pt above 30 GeV because of limited detector resolution." or similar

L205: add a comma after 1%

L237: is it a misprint "@NLO" -> "@LO" ?

Guillelmo Gomez:

- Type A (English) - l2 The CERN LHC has delivered large samples - l6. parton distribution function (PDF) - l9. imbalance [2] (too long separated) - l9. add a comma before "leaving" - l13. to the charm quark PDF, is - l15 [4], respectively - l16. pair of electrons or muons - l16-17. to have originated by a charm quark by applying charm tagging criteria [5] (c-tagged jet). - l18-19. to the generator level. - l21/22. add a comma before "as a function" and add a comma before "are" - l23. a c jet enriched [or a c-jet-enriched] sample - production where --> production, where - l16. to have originated with a charm quark->to have originated from a charm quark - l22. remove "the pT of the" - l114/115/117/... pt > X GeV check the distance between those words, they seem to be too large - l149. events as --> events, as - l150. backgrounds are --> backgrounds, are - l50. (light-flavor jet). - l55. In addition to theMG5_aMC samples, the SHERPA [11,12] event generator - l58. In addition to events with light-flavor and b jets, there is - l59. These samples were generated - l68-69. the NNDPF3.1 [26] PDF set, and GEANTfour [27] is used for the CMS detector response simulation. - l69-70. pp interaction in the same and nearby bunch crossings (pileup). - l104. nonlinear and nonuniform - l105. particle-level jets - l122. measured using the ``tag-and-probe" method - l126. and light-flavor jets - l128. and mistag rates for the b and light-flavors jets - l132. 1.2% for light-flavor and 20% for b jets - l140. as Z+light-flavor jet (Z+LF) events. - l152. subscript "q" in Roman - l153. Z+q jet [q in Roman] - l153-154. Sources of systematic uncertainties - l156. and the pT ranges in which - l189. in the c-tagged jet rats is estimated - l191. mistagging b and light-flavor jets - l195. the secondary-vertex mass SFbSFb and SFLFSFLF measurements. - l197. The JER uncertainty [can't start a sentence with an acronym!] - l198. coefficient in simulation up and down - l202. the identification (ID) and isolation (Iso) of electrons and muons are - l205. than 2 and 1\%, respectively. - l206. Top quark pair production cross section - l207. of top quark pair production - l217. The results from the electron and muon channels are \textsc{Convino} tool - l219. pileup, and top quark pair production cross section sources. - l220. and c jet, after the combination, are shown in Fig. 6. - l223-224. add spaces before every opening parenthesis and use "(theo)" - l228. using data collected by the CMS experiment at the LHC, corresponding to an integrated luminosity of 35.9 fb−1−1. - l232. were Z+light-flavor jet, Z+b jet, top quark pair, and - l233-234. for the signal with the Z boson pT<300pT<300 GeV - l234-235. add spaces before every opening parenthesis and use "(theo)" - l239. to extract the parton distribution function of the charm quark

Su Yong Choi:

L14: [3] and [4]. -> [3,4]. L17: "with a charm quark" => "from a charm quark" L27: "unfolded muon and electron channels distributions" => "unfolded distributions in muon and electron channels”

L55: [11][12] -> [11,12]

Li98: [31][32] -> [31,32]

L123: events, -> events.

L187: wrong reference [59] here.

Fig 2 and 3: add uncertainty band.

Fig 6: it’s hard to see the text in the legend.

L224: "MG5_aMC (NLO) predicted…" => “the value of 524 +- 11.7 (th) pb predicted by MG5_aMC (NLO)"

Burin Asavapibhop:

L7-11: Sentence "For example, ..... Into neutrinos" is too long. Propose to make 2nd sentence starting from ", one of the ...".

L47 : Is “MG5_aMC" official/general abbreviation?

L55 : MadGraph models -> MadGraph generator (models are misleading)

L83 : What is “better than” 7%? less/lower than? or up to?

L115-120 : two sentences are redundant, suggest to merge them and use bracket for electron channel instead.

L120 : m(ee) with small m

L123: Replace "," by "." at the end of sentence.

L159 : M(ee) with capital M — please fix for consistency

L128 : as you are talking about c tagging algorithm in previous sentence, I would propose to rephrase to

“The c tagging rate for c jets, and mistagging rates for b and light jets are measured using data events from W+jets, TTbar, and inclusive jet production processes, and compared to the simulation, where the reconstructed jet flavour is known from its hadron content.”

Figure2,3 : aren’t they Z+c-jets, Z+b-jets and Z+light jets processes? The legends on plots are misleading.

Figure2,3 : suggest to reorder the legend as same as the stack of histograms i.e. Z+c-jets, Z+b-jets, Z+light jets, Top and dibosons

Figure2,3 : please also add “(signal)” to the legend of Z+c-jets

L160 : shall this sentence about removing leptons from jet cone come since in the beginning L113-123 in Section4?

Jean-Baptiste Sauvan:

- l. 1: LHC delivers large samples -> probably this line can be rewritten in the form that such events have high cross section at LHC - l. 3: "c jets" -> "c~jets" (use a non-breaking space to avoid having it split on two lines) - l. 10: Z+c jet -> Z + c-jet (check the notation) (also please correct this in the rest of the paper) - l. 17: with a charm quark -> from a charm quark - l. 58: "As well as events with light and b jets": Rephrase, e.g. "In addition to events containing a Z boson with light and b jets" - l. 58: b jets -> b-jets (please check the notation) - l. 59: Samples of these -> These samples were - l. 60: “There is also background …. simulated using PYTHIA 8” -> “Background contribution from vector boson pair production is simulated using PYTHIA 8” - l. 83: “… resolution in the barrel is better than 7% …” -> "below 7%", or "smaller than 7%", or "lower than 7%" - l. 98: [31][32] -> [31,32] - l. 117: “ …. |eta| < 2.4. One muon must have p_T > 26 GeV.” -> can be rewritten as - “ …. |eta| < 2.4, with at least one muon with p_T > 26 GeV.” - l. 123: comma -> . - l. 124: "c tagging criteria" -> "c~tagging criteria" (use a non-breaking space to avoid having it split on two lines) - l. 148: Then the number …. -> The number …. - l. 149: c jet , b jet, light jet -> c-jet, b-jet, light-jet (please check the notations and correct in all the text) - l. 153: q jet -> q-jet - Fig. 2: Variable M_SV is not mentioned in the text. - Tab. 1-4 should be grouped together. Maybe a single table could be made since they are never referred to separately. - Fig. 3 should be closer to the text where it is referred - Fig. 4,5 should be closer to the text where it is referred - l. 163-164: Space missing before 30 GeV. - l. 172: Add TUnfold version - l. 218: "c tag/mistag" -> "c~tag/mistag" (use a non-breaking space to avoid having it split on two lines) - l. 219: “The values of the cross section …” -> "The measured cross section ….” - l. 221: "predictions from the generators" -> "predictions from the MC generators" - l. 223: “… pb, significantly lower than …” -> “ … pb, which is significantly lower than the MG5_aMC (NLO) predicted value of …” - Fig. 6: the legends are barely readable. Would be good to have bigger text. - Fig. 6: more details could be given in the caption, e.g. the MC predictions used. - l. 236: "The predictions from several generators" -> "The predictions from several MC generators"

Fabio Ravera:

English, Style and Formatting comments:

PDFTitle: Use sqrt(s) = 13 TeV as explained in the original SMP-19-011.tex file. Line 5: Should pT be (pT) Line 10: Remove “invisibly” Line 14: ..can be found in Refs. [3] and [4] or change line 44 so as not to use “Ref./Refs.” for explicit references. Line 17: “originated from a charm quark” sounds better than “originated with a charm quark”. Line 18: “observed data” could be replaced with simply “data” with little loss in meaning. Line 20: pp collisions are no time reference. Maybe say that data are pp collisions which were recorded by CMS … Line 30: 6 m -> 6\unit{m} so inserting a thin space (\,) between the number and the unit. Check other number-unit combinations in the paper to be sure that they either use newcommands (e.g., 13\TeV) or the \unit construction. Line 48: Remove the comma after “(DY) processes” Line 56: The line starting with “The value of the cross section” is already written in line 53. Should it be repeated? Line 60: “There is also background from vector boson pair production” does not sound nice, consider rephrasing it Line 70: …in the same or nearby bunch crossings. One needs to understand what “same” refers to but this is still better than “current”. Line 99: vectorial sum -> vector sum. ”vectorial” is not wrong; it’s just that “vector” is shorter. Line 108: is “degraded” the correct word? maybe smeared or the resolution is degraded Line 128: “quark jets”-> “light quark jets” Tab 1-4: All captions: “... the second IS the systematic...” or “...the second - the systematic...” Line 146: compared to -> compared with. Unless you really want to emphasize differences. Line 156: Here, and for the other tables, the caption should go above the table. Fig2: Reverse the order of the muon and electron plots. observed data -> data. compared to -> compared with. data is -> data are Fig 2&3: Would prefer that electron channel and muon channel labels are included on the plots Fig3: Reverse the order of the muon and electron plots. It might be better to start the jet pt distributions at 0 as well even if entries are restricted to pt > 30 GeV. Line 166: to a generator-level highest -> to THE generator-level highest Figs4&5: Reverse the muon/electron order. These plots are not very appealing esthetically and don’t seem to use the default TDR style (so are missing tick marks in two of the four sides of the frames). Line 215: Eq. (1) (Number in parentheses per the Guidelines) Line 216: The electron symbol should be in Roman font (not italic or Math mode). Tab 5: Use the same number of significant digits for a given column. The electron symbol (e) should be in Roman font per the Guidelines. Line 222: (LO)[,] and Sherpa (serial comma) cross section (no hyphen per Guidelines) Line 223: equals -> is Line 224: ..lower than [the] MG5_aMC (NLO) predicted value Fig 6: Use larger text in the legend of the upper plots. The ratio plots are too large compared with the main plots and the right plot exceeds the right margin. Unfortunately, where green hatch marks (for MG5_aMC) are overlaid on red (Sherpa), it looks brown (which is the color of the theory uncertainty). Line 233: cross-section -> cross section Line 241: Remove or fill in Acknowledgements

Kimmo Kallonen:

L8: SUSY abbreviation here is unnecessary, because it is not used elsewhere L175: Please change "c-jet" to "c jet" L223: Please change “, significantly lower” to “, which is significantly lower”

---

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

In general

(g) L54 and L57: (two places, as the "NLO" has been just introduced on L51)

"calculated to next-to-next-to-leading order with FEWZ" -->
"calculated to next-to-NLO with FEWZ"

Also, the last sentences of two paragraphs on L53-54 and L55-56 are

identical, I'm not sure whether the duplication can be fixed, e.g.

to change L55-56 from

"The value of the cross
section used is calculated to next-to-next-to-leading order with FEWZ." -->

(i) L81-85: (several places)

"results in a relative pT resolution for muons
with 20 < pT < 100 GeV of 1% in the barrel and 3% in the endcaps.
The pT resolution in the barrel is better than 7% for muons with pT up to 1

TeV [29]. In order to reduce the misidentification rate, muons
are required to be isolated. The isolation of muons is quantified as the
sum of the pT of PF candidates ..." -->

"results in a relative pT resolution for muons
with 20 < pT < 100 GeV of 1 (3)% in the barrel (endcaps),
and is better than 7% with pT up to 1
TeV [29] in the barrel. In order to reduce the misidentification rate, muons

are required to be isolated. The isolation of muons is quantified as the
pT sum of PF candidates ..."

(j) L87:
"less than 25% of the muon pT." -->
"< 25% of the muon pT."

(k) L94-96: (two places, the 1st of two "%"s can be removed)

"ranges from 1.7% to 4.5%. The dielectron mass resolution for Z -> ee decays

when both electrons are in the ECAL barrel is 1.9%, reducing to 2.9% when

both electrons are in the endcaps." -->

"ranges from 1.7 to 4.5%. The dielectron mass resolution for Z -> ee decays

when both electrons are in the ECAL barrel is 1.9%, reducing to 2.9% when
in the endcaps."

Also, the subject corresponding to the verb of "are" on L107 seems not very

clear and confusing with several commas on L106 in this long sentence, but

I'm not sure how to improve it yet.

(m) L113-118: (also, it'll be looked clearer if a comma is added after the
"in the muon channel" at the end of L115)

"Events are selected online with
a single muon trigger requiring at least one muon
candidate with pT > 24 GeV (muon channel), or
a single electron trigger requiring at least one electron
candidate with pT > 27 GeV (electron channel). Offline, in the
muon channel two opposite-sign muons satisfying identification and isolation

criteria are
required with pT > 10 GeV and |eta| < 2.4. One muon must have pT > 26 GeV. In

the electron channel, the offline selection requires two opposite-sign
electrons with pT > 10 GeV and |eta| < 2.4 ." -->

"Events are selected online with
a single muon (electron) trigger requiring at least one muon (electron)

candidate with pT > 24 (27) GeV for muon (electron) channel. Offline, in the

muon channel, two opposite-sign muons satisfying identification and isolation

criteria are
required with pT > 10 GeV and |eta| < 2.4. One muon must have pT > 26 GeV. In

the electron channel, the offline selection requires two opposite-sign
electrons with the pT and |eta| ranges same as muon."

(n) L126:
"of c jets from b jets and light jets," -->
"of c jets from b and light jets,"

(o) L149:
"for the Z+b jet and Z+light jet" -->
"for the Z+b and Z+light jet"

(d) L120: (then can be shortened by using the "mZ" correspondingly)

"is required to be close to the mass of the Z boson," -->
"is required to be close to the mZ,"

(6) L124-125, L128 and L131. It seems looked much better if a hyphen is added

in the adjective "c tagging", e.g.

L124-125:
"satisfying tight c tagging criteria ..." -->
"satisfying tight c-tagging criteria ..."

L128 and L131 are similar.

(7) L127, the expression of
"a b hadron" is looked quite odd similar as
"a b c d xxxx".

It may be improved by changing from
"a b hadron" -->
"a bottom hadron"

(10) Tables 3-4, two captions are almost identical except differing by only

one word at the ends of 1st lines; thus two Tables may be combined with an

extended 1st line of Table 3's caption, i.e.

"Table 3: Values of Z+light jet, Z+c jet, and Z+b jet scale factors measured

in the muon channel," -->
"Table 3: Values of Z+light jet, Z+c jet, and Z+b jet scale factors measured

in the muon (upper Table) and electron (lower Table) channels,"

Also, as the header rows of two Tables are identical, when combining two

Tables, the header row of Table 4 can be replaced by a double border line.

(11) L200, L204-205, L208 and L210. Those systematic uncertainties

percentages (e.g. 4.6%, 5%, 2%, 1% and 2.5%, etc.) may should cite some

reference articles, so that readers would not wonder why these percentages

are chosen, why not any other arbitrary percentage numbers.

At least, the luminosity uncertainty of 2.5% should be given a Reference, to

be consistent with all other CMS papers, i.e.

L210:
"+-2.5%." -->
"+-2.5% [xx]."

(12) L203-205 and L222-223 can be shortened from

(a) L203-205: (three places; also, it may sound better if two commas are

added after the 1st two words and before the last word of

sentence)

"For electrons the uncertainty is less than 5%, while for muons uncertainties

for identification and isolation are less than 2% and 1% respectively." -->

"For electrons, the uncertainty is < 5%, while for muons uncertainties
for identification and isolation are < 2 and 1%, respectively."

(b) L222-223: (two places)

"cross-section value for Z pT < 300 GeV equals 413.5 ..." -->
"cross-section for Z pT < 300 GeV is 413.5 ..."

All suggestions above have been carefully considered. In some cases they have been superseded by larger changes, in some cases they have been applied, and in some cases we prefer to keep the original.

---

end --------------------------

- Title: of Z bosons -> of a Z boson Fixed

- Title: "Z bosons and charm jets" doesn't seem to imply associate production, propose to change to "a Z boson with charm jets", or "associated production of a Z boson and charm jets" Fixed

- lines 5-12: The physics motivation that you expand in the text are SUSY searches of the stop quark. Since now most of the phase space accessible at the LHC for SUSY is excluded, you should expand (e.g. adding some reference) the other two motivations: c-quark PDF and (QCD?) theoretical models. Removed "main background" for susy, added "...test the possibility of observing of intrinsic charm quark component in the nucleon..."

- Figure 1: the arrow of the virtual fermion is pointing in the wrong direction (unless this is meant to be a c-bar) Fixed

- l 61: You miss the contribution of misidentified leptons. Even though small, it should be cited at least, to say that it is negligible and neglected (if it is), since most of the Z->ll analyses consider it. As I can see in other papers, misidentificaion coming from Wjets and QCD multijets are not considered. Contributions from misidentified leptons are negligible

- l 63 vs l 68: it is confusing since L63 says that NNPDF 2.3 is used, while L68 says that "Samples are generated using the NNPDF 3.1 [26] PDF set" !!!!! Yes, this is an error: ttbar uses NNPDF 2.3, while Madgraph LO, NLO and Sherpa use NNPDF 3.0 (not 3.1). Fixed

- ll 82-83: "The pT resolution in the barrel is better than 7% for muons with pT up to 1 TeV [29].” Is this really relevant for this analysis? The bulk of events has pT(l)<100 GeV, so we suggest to remove this sentence. Fixed

- ll 88-96 (and Table 5). From the description, and the systematics table, it seems that there is no isolation for electrons. Why is this the case? All the EGamma WPs require isolation. If this is not applied, then the concern about the presence of fake background is even more important. If a standard WP of EGamma is used, inluding isolation, then you should
• add the isolation in the description of the electron selection
• specify in Table 5 that the systematics is for the combination of ID + isolation Yes, isolation criteria is included to electron ID. Isolation and id for muons are done separately, this leads to lots of questions, so id and isolation were combined (mention it in tables as Id/Isolation). For electrons this is already combined, for muons we can sum them and present as Id/Iso

- l 101: "tracks" -> "charged particle candidates" Fixed

- ll 113-115. Why the single-lepton triggers have been used, instead of the double-lepton triggers? Especially for the electrons, the real pT threshold on the single electron is not, as you write, 27 GeV, but a variable threshold ranging from 45 to 25ish GeV, which is the L1 threshold. During 2016 that was higher than the nominal HLT threshold (27 GeV). So at least you should remove the 27 GeV, and say that a higher threshold was used (the exact value varied during the LHC fills). But for the electron channel using the DoubleElectron HLT path would have decreased a lot the threshold and gained statistics. Also, the offline threshold pT>29 GeV on the leading electron is below the L1 threshold of the single-electron trigger (L29). Have you modelled well the trigger efficiency scale factor vs pT? Maximum of leading lepton pt is higher than trigger pt threshold, so there is no gain in statistics from lowering pt threshold. On the other hand it is easier to calculate trigger SFs for Single lepton trigger.

- ll 115 and 118: do you really need the requirement of the two leptons being opposite sign? Since the combinatorial background is very small, especially in the electron case you could have avoided that. It is standard selection for processes with Z+smth

- l 120 : "71 < m ee/mumu < 111 GeV "-> "(71 < M(ee/mumu) < 111) GeV or 71 GeV < m(ee/mumu) <111 GeV " Fixed

- l 121: “small residual differences in the trigger”. How small are they? See AN-2017/340 v17, Fig.25. At pT=30 GeV the electron trigger SF average in 2016 is 80% in the central barrel and down to less than 50% in the second half of the endcap. Please justify better this. Please take notice, that 50% is not SF, but efficiency in data. Efficiency can be low, especially for low pt or high eta, but the difference between data and MC is defined by ratios of efficeincies in data and MC, which is much closer to 1.

- l 142: "deltaR < 0.1" -> "deltaR = 0.1" Fixed

- ll 144-145 and Fig. 2.
• If only the invariant mass of the secondary vertex is used to discriminate S from B, and the Z+jets events are only selected with the loose cut on 71<m(ll)<111 GeV, then the background from fakes must be non-negligible. Can you comment at least on a systematic added about its normalization?
• The m(SV) does not seem to be very discriminant between the light/c-/b- jets. The shape seems to be very similar for the 3 types of quarks. Since this is the focus of the paper, this choice should be better motivated.
• We added comment, that SVM allows to discriminate flavor components, despite it was used by c-tagger.

Related to this: with so low-discrimination, have you checked that you can obtain a good fit also with very different combination of the 3 SFs (light- ,b- and c- jets)? I.e., have you checked that you have not multiple minima in the likelihood?) 1)According to other Z+jets analysis fakes from W+jets and QCD are negligible, other backgrounds from top and dibosons are taken into account. 2) Indeed separation of different flavors by SVM after c-tagging is much worse, than before c-tagging, correlation between fit results SFb and SFc is ~-0.8. To check large variety of possible SFc, the fit was done in range (0,2) for all SFs (SFl, SFc and SFb). Negative SFs and SFs > 2 look unphysical, so if there are any other solutions, they are not taken into account, within determined range (0,2) the solution corresponds to maximum of likelihood.

- l 150: "by fitting the secondary vertex mass distribution": much more details are needed to explain the procedure.
Is it a template fit? are the templates taken from MC? do these different processes have different-enough mass shapes to be able to discriminate? maybe adding a figure of the templates (normalized to the same area) could help We added, that despite the fact, that SVM is used by c-tagger, it still allows to discriminate flavor components

- l 151: "fits are performed separately": does this really mean separately? so the scale factors obtained in the Z-pt bins
are completely independent from those obtained in the cjet-pt bins? Yes, these fits are done separately for Z or c-tagged jet pt bins. A good cross-check is that light, charm or bottom pt integrals after such fits are close (within max 4%) for both variables.%END%

- lines 148-156: the text should clarify the role that these scale factors have in the analysis, and in particular how they affect the cross-section measurement Fixed: The number of \zcjet events is calculated as product of the number of \zcjet events, predicted by MC and the normalizations for the \zcjet component.%END%

- tables 1 and 2: replace "pT interval" with "c-tagged jet pT" Fixed

- tables 3 and 4: replace "pT interval" with "Z candidate pT" Fixed

- tables 1-4.
• Why you want to show 4 tables of scale factors for (c-jet pT, Z pT) x (ele,mu)? You could show just one example for c-jet pT and Z pT, choosing one channel, since in principle they should be independent on the Z final state.
• since you show them for the 2 lepton flavors, large difference of a SF in the two channels (e.g. 21% difference for SF_c for pT(c) in [90-250] GeV, but there are many cases) could be an indication of the interchangeability of the three components to have a good fit. The systematic uncertainty is large 5-15% depending on the bin, but does it indicate that you don’t have sensitivity to discriminate the 3 flavors? The main goal is to discriminate charm component from light and bottom, and since the descrimination is poor, light and bottom are anti-correlated,which can lead to big difference. We have also checked that fixing light component doesn't have much impact on the final result. As a good check how large the difference is, the integrals for each components were compared for Z pt and c-tagged jet pt.

- all figures: there are a number of cosmetic improvements that could be done. for example: increase the font size for all labels
(particularly needed for the legend and the CMS label); "DATA/MC" -> "Data / MC"; use consistent capitalization in the legend
(why is "light" lower-case while "Top" is not?); also in the legend replace "Observed -> Data" Changed, light, char, top etc changed to Upper case.

- table 2: why are the uncertainties larger than table 1? Despite the fact, that we change the parameters to measure systematic, the result depends on statistics, so that systematic includes stat fluctuation. The bigger stat uncertainty (electron channel - table 2) the larger the uncertainties

- l 159: same as line 120 Fixed

- l 160: "within deltaR < 0.4" -> "within deltaR = 0.4" or "in deltaR < 0.4" Fixed

- l 163: I would rephase: "primarily events with c jets with p T generated at < 30 GeV but reconstructed at > 30 GeV because of detector resolution"

- l 164: "because of detector resolution" -> "because of finite detector resolution" or "because of detector resolution effects" Fixed

- l 166: "within deltaR < 0.3" -> "within deltaR = 0.3" or "in deltaR < 0.3" Fixed

- ll 182-185: you should say that you don’t consider the anti-correlated combination of muR(up) and muF(down), according to the usual prescription of this systematic. Fixed

- l 209: add a reference on the luminosity uncertainty. Fixed

- Fig. 4 and 5: since the behavior of the purity as a function both pTZ and pt(c-jet) is determined by the jet pT resolution (since the Z->ll resolution is very good for both ele and mu), why do you need to show both? The two plots match almost by definition (because the Z pT balances the c-jet pT). So you could show only the Z pT one, since it goes lower in pT, and beyond 40 GeV should be the same (it cannot be said from the plots, since the binning is different, but it seems so).

Other plots are shown separately for muons and electrons. In this case the difference is negligible, but maybe it is worth showing to the reader, so that it would be clear, which steps of the unfolding are different.

- l 213: maybe remove "Eq.1" from the text Fixed

- table 5: why some of the jet-related systematics are so different for the ele/mu cases? E.g. JER +0.5% (ptZ, mu), 1.2% (ptZ, ele) There is much difference between two channels: Z pt and as a result jet pt spectra are different because of differnt lepton selection efficiency. Second reason is that uncertainty depends on statistical fluctuations, which can also lead to difference between channels.

- l 224: you measure a xsec that in principle is known at NLO, and it disagrees for more than 20% wrt the predictions.
Can you add a physics comment about the importance of this experimental input to the theory?
Also, since you justified this measurement with stop SUSY searches, can you comment on the impact of the precision of your measurement on those searches? We decided to add some interpritation: e.g. since there is good agreement with no c-tagging for NLO, it may indicate that pdf of charm quark can be tuned.

- l 231: same as line 120. Changed ?

Type B

• L3: “hadronic jets” -> “jets”, consider to drop the superfluous “hadronic”

Fixed

• L7-L11: Please consider rephrasing this, as it may be difficult to read

Fixed

• L15-L17: Please change “Here we present a study of events containing a candidate Z boson decaying to a pair of electrons or a pair of muons, and at least one jet identified as being likely to have originated with a charm quark (c-tagged jet) by applying charm tagging criteria [5].” to “Here we present a study of events containing a candidate Z boson and a candidate c jet. The Z boson decays into a pair of electrons or muons, and the c jet is required to pass charm tagging criteria [5].”

FIXED

• L27: What distributions? Are you still referring to the SV mass distributions or the unfolded differential (in pt) cross-sections?

Fixed

• L46-L70: Please consider putting this information in an easier-to-read table.

We believe it is better to keep the plain text with all the details, instead of adding a table.

• L50: Are you using MLM matching with the NLO sample?

MLM can't be used with NLO

• L97: “Hadronic jets” -> “Jets” (see comment for L3)

Fixed

• L113 onwards : Please mention that the HLT objects are matched to the offline objects (if they are).

They are not matched, we use combinatorical formula for efficiencies in data and MC

• L200: Is there a proper reference (besides the TWiki-page PileupJSONFileforData) for the 4.6%? Please add.

Fixed

• L201-203: Here you also mention the “isolation of… electrons”, but you neither define isolation criteria for electrons in section 4, nor is an uncertainty on electron isolation quoted in table 5. Please clarify the situation about the electron isolation and associated uncertainties in the text.

Yes, isolation for electrons is included in their identification, while for muons these two criteria are separate. We propose to combine Iso and Id for muons, and mention Id/Iso for both muons and electrons.

• L201: Please mention the trigger efficiency scale factors here too

Trigger efficiency scale factors errors were neglected, as the result uncertanty is smaller than 0.5%

• L239: Please change “These results can help to fit the PDF…” to “These results help with fitting the PDF...”, as you should be more assertive here
• Fig. 1: Please redo arrow direction for the quark (flipped in the middle)

Done!

• Figs. 2-3: No error bars visible in the stacked plots: Uncertainties so small that they are invisible or are they not plotted? Please clarify.

Only data stat errors were plotted

Strategy, paper structure, emphasis, additions/subtractions, etc:

• Abstract: The Abstract should include a sentence indicating the relevance of the results - are they the first at 13 TeV? Do they significantly improve on past results? Also, it is not clear why the fiducial cross section measurement is mentioned at the end. (The title refers to differential cross sections without fiducial restrictions.)

Fixed, maybe we should add comparison with 8 TeV result in the summary.

• Line 5: “Test existing theoretical models…”. It is a bit too vague, there are many theoretical models including BSM ones. I think you mean to test SM predictions.

Maybe there are some other models beyond SM predicting Z+c, e.g. some extentions. We can't prohibit to fix those models using this data.

• Fig.1: The arrow along the quark line always has to point in the same direction, i.e. the middle arrow has to be reversed

Fixed

• Line 18, 26: “unfold the data to generator level” - the data are unfolded to “truth” (or model) and a generator is often used as a proxy but saying that data is unfolded to generator level is not correct; You should also cite something for Unfolding (https://www.researchgate.net/publication/309738352_Data_Unfolding_Methods_in_High_Energy_Physics or else). Similar for line 26, because the ultimate goal of the real data measurement is not to describe “generator level”; It is worth rephrasing this part

We added definition, what we call generator level.

• Line 56: Is the background Z+n jets with n>4 determined to be negligible?

One can see from SMP-20-009 , that Z+3,4... / Z+>=1 ~ 6%, so this process can be neglected.

• Line 59: Single top production is mentioned as a background here but it is not clear if this is used in the analysis. For example, line 127 does not include single top in measuring tagging and mistagging rates. In figures 2 and 3 it is not clear that single top is part of the “top” in “Top and dibosons”.

Yes, both single top and ttbar are considered in the analysis.

• Line 68: Specify the QCD order of the NNPDF 3.1 set. NNLO?

Fixed, version 3.1 was a mispell, changed to 3.0

• Line 76: Since the Abstract mentions “electrons or muons” and this is the usual order, you should keep this order whenever the two are paired together. So describe electron reconstruction first.

Fixed

• Line 78: The “primary vertex” has not been defined. If the standard definition (highest summed pt-squared of charged tracks) is used then the paper should clarify whether or not the leptons and jet are required to be associated with the PV.

Fixed

• Line 113: Describe the electron trigger first - before the muon trigger.

Fixed

• Line 117: “One muon...” (similar for electrons on the next line ) - and no requirement this to be the triggering muon? More generally, it is not specified if trigger matching is applied, which you need if using T&P trigger SFs.

Changed to matching trigger object with highest pt lepton, same result, as for combinatoric formula .

• Line 129: “where the reconstructed jet flavor is known from its hadron content”. In simulation you could trace the jet to a parent quark or gluon, so this could be more simply “where the reconstructed jet flavor is known”.

This phrase can be crucial, because there are two flavor definitions of flavor in CMS: hadron and parton based. We use the first variant.

• Line 131: Since the c tagging efficiency and associated mistag rates are crucially important for this measurement, it would be useful to indicate the approximate uncertainties on these numbers

Fixed, but more details on scale factors are not discussed in uncertainties section. There are general information, that SFs are varied within 1 std

• Line 137-140: It may be useful to specify what happens if there are both b and c jets above 10 GeV (possibly the c jet much more energetic)

This will be a repeating: if there is one b-jet with pt > 10 GeV, the event is bottom, regardless what is the pt and flavor of other jets.

• Line 140-142: Why do you have to correct gen-level leptons for photon radiation and not just use the momentum before any radiation (or after radiation depending on what you mean)? Also, it is not totally clear what is the message this paragraph should pass. Please, clarify if it is Final State Radiation you are talking about or any radiation.

At this generator stage leptons are simulated with emitted photons, so to take them into account, their is a correction (dressed leptons)

• Line 144: A short description of how invariant mass of tracks used as secondary vertex mass is needed here

Added short description how SVM is used.

• Line 151: Are the fits done separately? Why not simultaneously across Z pT bins for each c-tagged jet pT bin and simultaneously across c-tagged jet pT for each given Z pT bin?

It is not possible to split events simultaneously into Z and c-tagged jet pt bins: there wouldn't be enough statistics in each bin to perform SVM fit. In addtion it wouldn't be differential cross-section, but rather double differential cross-section measurement.

• Tab.1-4: How are these used? E.g. Tabs. 1/3 and 2/4 are different parameterization of the same events - which one do you apply? Would a 2D SF make more sense? Also, are the differences between the muon and electron channels expected? We are wondering if these tables are not too detailed given the fact the other SFs are not discussed in the paper. Please, consider to draw all of them on a single (or couple of) plot(s).

We decided to keep present tables and text. SFs for electron and muons are independent measurements of Z+c and as such there can be fluctuations. Correlations among them are small, therefore we can't really comment on the difference between them.

• Line 154: The sentence says the source of uncertainties for SF will be discussed in Section 6 but this section does not mention anything about uncertanties of SFs.

It doesn't say, that uncertainties for SFs are discussed in section 6, but sources of uncertainties in general.

• Line 158: The gen level is 26 GeV and your reconstruction cut for electrons is 29 GeV. It is worth clarifying in the paper how large an effect is that (effectively a small acceptance correction), given electrons and muons are treated differently.

Fixed: added sentence, that acceptance takes into account correction for gen-reco cuts difference.

• Line 164-167: This sentence is a bit too long and makes it confusing; 1) “The fraction… is estimated from…” doesn’t clearly describe how it is estimated. In the caption of Fig. 4 you state what you plot “Fraction of selected events originating...” which appear clearer; please, follow what written in the caption. 2) You should specify if the lepton pair is matched or not; so follow the caption but split this sentence to add additional information. 3) As you are talking about “reconstructed” and also about “generated” quantities, clarify in the caption of Fig 4 what is on the X-axes (your caption implies “reconstructed” but “generated” may be a better choice)

• Line 169: Somewhere in the text, not necessarily here but it is an option, you should clarify what is your “true” distribution (or your “generator level” as you call it). One of the most important points is to clarify if your “true” includes or not Final State Radiation, it makes a difference for theoreticians (electron and muon response matrices, at least for Z-boson pt, should be sizably different, because reconstructed electrons include rediated photons and reconstructed muons do not).

Changed the sentence to "The generator-level leptons are dressed by adding the momenta of all photons within $\Delta R = 0.1 $ around the lepton directions. ".

• Line 173: It should be “efficiency” of something, probably “efficiency of event reconstruction” or “event reconstruction efficiency”, don’t use “(...) “ to define it. It is better to somehow relate it to fig 4 because the two together are most interesting and thus better have the same X-axes (now the implication for fig 5 is that you plot “generated” quantities). Please, consider swapping fig 5 and 4, showing first that you get ~20% of the actual events and only then that the fiducial corrections are small at high pT. Then the response matrices as the last piece.

• Line 175: ..for electron and muon channels. Invert the order of Tables 3 and 4.

Fixed

• Line 177: Theoretical and model uncertainties should be present also in the theoretical models, so theoretical/model points on plots should have associated uncertainties. While there are some uncertainties on the figures it is not clear what those are. The PDF uncertainties seem to be playing a prominent role in your experimental measurements themselves too.

• Line 175: If c tagging is the dominant source of signal loss, why is the muon and electron channels so different in total efficiency (by about x2)? Please specify in the text.

Added, that c-tagging and lepton selection efficiencies contribute to acceptance

• Line 187: The reference number has jumped to 59 - although the last entry appears to be 39. It may be unclear for the reader how the 90% uncertainty level is “scaled” to 68.3%

That's an error citation. Changed the whole phrase for a more commone one.

• Line 215: What is the value of BR(Z->ll) used?

Fixed

• Results: A description of how uncertainties of theoretical results are calculated is needed. Moreover, why are the uncertainties of MG5_aMC LO and Sherpa very small compared to MG5_aMC NLO?. Fig. 6 caption should explain whether “Total” uncertainty includes “Stat” or not?

PDF and QCD uncertainties considered only for MG NLO. In new version PDF uncertainties added for Sherpa.

• Line 218: “their correlation” is an important point. Please, consider to specifying what is correlated, how you estimate it, how important the overall effect is

correlated sources listed, added, that correlation was assumed equal to 1.

• Line 222-224: This sentence risks to lessen the results achieved in your analysis. Since it is a major achievement and it is not stressed enough in the conclusions, in this part you should elaborate what is only cited in the last two sentences of the summary.
• Line 239: It is not clear what “PDF of the charm quark” refers to - is this a reference to the c-quark component of the sea quark portion of parton distribution function sets? If this idea is important it should have been developed in the Introduction. We changed to more general sentence, that it can imporove constraints on parton distribution function of charm quark.

• Summary: The context of the results should be given here (and in the Abstract). For example, is this the first measurement at 13 TeV? Added, that it is first measurement at 13 tev

• Reference [35]: would JME-17-001 fit better? Fixed
• Reference [36]: would JME-18-001 fit better?Fixed

## Type B
- Abstract: Add the fiducial region used and the measured value of the inclusive cross-section.
Fixed

- l. 5: "can test existing theoretical models": which theoretical models are referred to here? At least some examples should be given
We present list of MC event generators (MG LO/NLO and sherpa) later

- l. 18,19: "we unfold the data to generator level". The term "generator level" is too vague. It is for instance better described in [4]
We clarified in the beginning, what we call generator level.

- A short discussion on the backgrounds could be added in the introduction before the "simulated samples" section

As I can see in other papers, it is usual to discuss backgrounds in simulated samples section and skip this information in the introduction. (see e.g. SMP-19-004)

- l. 50: is the NNLO inclusive cross section rescaling used here as in the other MCs? It seems strange that this one is not rescaled while the other two are rescaled to NNLO
The MC weights are +/- 1 for central values, and some weights, different from 1 for qcd/pdf variations. These don't take into account cross-section values, leaving the choice to the user (at least as I understood)

- l. 53 and 56: "The value of the cross section..." -> "The value of the inclusive cross section..." (to make it clearer that the NNLO cross section is inclusive in the number of jets, >=0)
CHanged to inclusive cross-section

- l. 118: mention identification criteria, as for muons
Fixed

- l. 135: Are jets really made from final state hadrons only? This should rather be all visible final state particles.
changed to "all stable particles resulting from hadronization"

- l. 158: "(at least one with p T > 26 GeV)". This is not consistent with the reco level selection in the electron channel (29 GeV). It means that there is an extrapolation to a wider fiducial region in the electron channel. That should be more explicitely written (in particular this is visible in Fig. 5, with lower efficiency for electrons, but no comment is made). This kind of extrapolation can also be subject to uncertainties, which would affect the efficiencies shown in fig. 5. Though it seems not uncertainty is associated to that.
The acceptance takes into account this extrapolation to wider regio, just as it is done with c-tagging efficiency.

- Fig. 2,3: systematic uncertainty bands should be added.
Fixed

- Tab. 2-4: It is not clear whether these scale factors uncertainties take into account differences from different MC generators used to extract the relative fractions of signal and backgrounds. The fit may give different values for the numbers of signal events using different MCs. But it is not mentioned anywhere. If it has no impact, it should also be mentioned.
SFs depend on MC choice, so it is not correct to compare them, but the product of SF and prediction of MC is measured number of charm events, which is compatible within stat errors for LO and NLO.

- Fig. 4,5: it seems that no uncertainties are associated to these values. While they should depend on the choice of MC used to derive them. Something related to that should be mentioned in the text.
We have checked, that unfolding of measured charm events using LO responce, bkg and acceptance is compatibles with baseline NLO values.

- Fig. 4,5: Comments on the binning choice should be made. In particular it seems that the same binning (apart from the different starting point) has been used for pT(Z) and pT(c), while it is expected that the migration matrices will be very different for pT(Z) and pT(c) (lepton vs jet-based quantities), and would require different optimal binning for the unfolding.
The binning at reco level was chosen so that number of events was ~the same in each bin and it was possible to make a SVM fit.

- Uncertainties on the purity, efficiency and unfolding related to the choice of MC and to the choice of the unfolding method should be mentioned (or it should be mentioned if they are negligible). Constructing the response matrix and deriving purity and efficiencies relies heavily on simulation. Different MC might give different results and this comparison should be done and mentioned. Also inverting the response matrix can be done in many different ways, and the unfolding methods have in general parameters that need to be tuned; which means that there can be a dependency of the unfolded results on the way the unfolding has been done. Unfortunately no discussion at all is presented here, and it would be good to have some details on the various checks done related to the unfolding.
We have checked, that unfolding using sets of bkg, acceptance and responce from different generators (LO and NLO madgraph samples) provide compatible results (within stat errors)

- l. 173-176: The differences between the electron and muon channels should be explained. Added, that the loss is due to ctagging efficiency AND lepton id efficiency, this should explain the source of difference

Section3 : which PDs have been using in the analysis? It seems missing from section 3, only MC is given. Fixed

L14: Why do you use Z+b as a reference for 7 TeV, [3] The main goal is Z+b, but since Z+c is one of the main backgrounds it is also studied in this work

L79 : Impact parameter is a jargon, please give some clearer physics meaning It is a jargon, yet it is used in almost all papers, so chosing another term can onle mislead the reader

L120: How is the mass range of Z chosen? It is standard choice for Z+jets analysis

L144-145 : are "the invariant mass of tracks associated with the secondary vertex” and "the secondary vertex mass” same thing? I would guess they are the same thing but it’s not so clear to reader. Changed everywhere to M_{SV}

L144-147: It is unclear how Fig 2 is used to discriminate between signal and background. Could you please clarify? It was added to the paper, that different flavor components of DY have different shapes of SVM distribution, so this is used in the fit

L150 : it is not clear from the text how the fitting is done to get normalisation of each Z+q jet processes. It is maximum likelihood template fit. Added this to the text

L158 : Why the electron pT is 26 GeV for simulation but 29 GeV for Data? The difference is for detector level and generator level, not data and MC. 29 GeV cut is applied at detector level because it has to be greater than trigger threshold. On the otherhand, we wanted to keep the same generator level signal definition for both channels, so 26 GeV is the same cut for electrons and muons at gen level. The difference between 29 and 26 for leading electron is small, but taken into account in unfolding

L189: c-tagging efficiency, is it recommended by JetMET group or you derive it? If from recommendation, please provide a reference. Same comments for JES/JER, Lepton ID, Luminosity. These tag/mistag SFs and their uncertainties were recommended by BtagPog. They are stored only on twiki, so there can't be reference for them

Summary : do we know why the LO shows better agreement to data results than NLO? One of possible explaination, is that in both MC LO and NLO charm component is overestimated, but LO predicts less Z+jets, than NLO, which has good agreement with data for Z+jets. SO these two effects compensate each other.

* Type B ***

L11: are you sure that the search for t~ -> c + LSP is limited by the Z+c production cross section? The largest uncertainty on the Z->vv prediction is the size of the dilepton control sample, so it’s not clear how a better measurement of Z+c in visible decay modes helps the search. Please clarify.

Left the reference but change the text to make a more general statement that Z+c is an important background for new physics searches with a charm and missing energy in the final state.

L96: "reducing to 2.9%" => why is it reducing? 1.9% to 2.9% is not reducing. Please clarify.

fixed to "...degrading to..."

L148: please justify this approach. What is the level of your trust?
Mismodelling of dibosons and single top can be neglected: as sum of these components is ~1% of DY events. From varyng normalization of ttbar by 10% we see that contribution to result uncertanty is smaller than 1%

L148-150:
- where is the template taken from? If they are from MC, what is the uncertainty on the shape?
- according to Fig 2, the M_SV shapes look similar (maybe it’s because the templates are stacked…). So, it’s not clear how the similar-shape templates are constrained by the fit. Please clarify.
The templates were taken from MC. In order to validate, how MC describes shape of SVM after applying c-tagger, it was compared with data in ttbar and W+c events. Though there is small mismodelling of SVM shape, the difference is small comparing to other errors. The shape difference after applying c-tagger is smaller, than without any HF taggers, but b-jets tend to have larger tails for larger SVM values, which allows to discriminate charm from bottom. The anti-correlation between charm and bottom components is ~-0.7/-0.8
L175-176: if the dominant effect is due to c-tagging efficiency, why are e/mu numbers so different? I guess it’s due to lepton efficiency and you might want to mention that.
Fixed

Fig 3: the first bins of p^{ee/mm}_T distributions show discrepancy of 10%. Is this understood?

It looks like the residual mismodelling of Z pt spectra: in case, when no c-tagger is applied the "wave" in data and MC ratio is observed for small Z pt values. This plots are shown in the AN for both ee and mumu cases.

- Type B (physics, clarity)
- abstract. in the fiducial --> in a fiducial (since there are many possible fiducial regions, you just use one)
Fixed

- l5. can test --> test
Fixed

- l13. We would remove "sensitive to charm quark PDFs", it doesn't add anything

The sentence was changed, though we still mention, that the study can provide some information about c quark pdf.

- l16. Since the whole paper is muons/electrons, we suggest to swap it here too. Make sure you are consistent in the whole paper
Whole paper now follows the electron/muon order

- l20. The data used, corresponding to an integrated luminosity of 35.9fb-1 at sqrts - 13 TeV, were recorded by the CMS experiment in 2016
Fixed

- l25. The secondary mass hasn't been defined, what is it? you need to explain it
Fixed: removed.

- l28. physics generators --> event generators

Fixed

- l28. and to compare with predictions->and to compare them with predictions

Fixed

- l48. To which oder do you generate Z+c?

Fixed

- l53. We assume you use FEWZ to obtain the full inclusive generator, and then the split in nJets is assumed by the generator, right? say it

FIxed, added : "...for inclusive Z + jets process."

- l55. You need to mention the SHERPA version

Fixed

- l60. You need to mention the POWHEG version

Fixed

- l61. You need to mention the PYTHIA version

fixed.

- l65. Do we really use 19?

we decided to keep it

- l70. We think it would be good to add the average pileup in data in 2016

added , which in 2016 equals on average 23.

- l72. The particle --> The CMS particle (notice there are many possible PF algorithms, that's why you need to define as the "CMS" one)
- l77. You should add a reference here We used standard sentence about muons in CMS

- l78. How is the PV defined?
Fixed

- l83. Actually, there are very few "fake" muons, but rather nonprompt muons. It's better to write it in that way

- l86. It is not clear how the pileup is substracted by just reading the text.
Its a standard phrase in all papers about correcting isolation parameter.

- l88-96. We are surprise there is no mention about electron isolation, we assume you use it, don't you?
Added text about electron isolation

- l97 and that paragraph. Are you applying a lepton-jet overlap removal? this is not mentioned anywhere We used standard description of jets in CMS

- l142. Shouldn't you mention this is usually call dressed-leptons?
We don't mention dressed leptons in other parts of text, so there is no need to inroduce this standart term - dressed leptons

- Tables 1 and 2 (major point). We don't understand the results, and why you use two different SFs for muons and electrons. There is no reason why the jet SFs should be different for the muon and electron channels, did you check it in MC? do you see different performance between them? it should not be the case. In particular, there are some bins with significant differences in the SFs, is it understood? For instance (but not only) pt30-37 SFc (0.85 vs. 0.69)
We don't understand the difference between two channels: there is large statistical uncertainty for these values, within which they overlap. The same difference was observed in other analysis, e.g. SMP-19-004.

- Figs. 2 and 3. Could you make slightly larger legends?
Fixed

- Fig.2 Muon and electron channels are not indicated in the two figures
Fixed

- Tables 3 and 4. Same comment as before, we are not expecting sizable differences. In particular, the SFl trends for muons and electrons are exactly the opposite, is it understood?
It is not well understood, since separation between flavor components after applying b-tag is small, the uncertainties are large, especially for light component, which is the smallest comparing Z+b and Z+c.

- l170. What is the default MC used for the response matrix for the unfolding?
MG NLO, added to the text. We have also cross-checked using LO, since bkg, acc, and response shouldn't depend on generator choice.

- l185. Do we assume you don't take the extreme cases? say it!
Fixed

- l193. Can you quote the rough input uncertainties?
Added SFs variation estimates

- l198. Can you quite the rough input uncertainties?
Added SFs variation estimates

- l199. Add a reference about it
Fixed

- l204. The 5% electron uncertainty is a very large overestimation, how did you get that?
Yes, this is too conservative upper limit, changed to 3%.

- l208. Can you justify that value? is there any reference?
Yes, reference was added. This approach was used in other analysis with Z boson and jets. - l210. Give a reference about the luminosity
Fixed - Fig.6 Z and c jet is not indicated in the two figures

- Section 7. In general, we don't understand why you split the muon and electron channels, and then you perform a combination afterwards. The jet results should not depend on the lepton flavor. Nevertheless, since you do so, you must report both results, and their compatibility

It is not clear, how to select muons and electrons simultaneously (there are samples enriched with either muons or electrons). This apprach is used in other studies, which include Z boson.

- Table 5. You need to add two significant digits in several cases (e.g. 4 --> 4.0, 0.4 --> 0.40). In addition, you are missing the luminosity uncertainty (constant, maybe mention it in the caption?), but more importantly the total uncertainty should be quoted) We quote total theoretical and experimental uncertainty in Results and Summary sections, after combining two channels

Unlike other uncertainties, we use uncertainty on luinosity after unfolding, when unfolded distributions are devided by total luminosity. This leds to simple varying within 2.5% so it was not added to the table.

- l172. Which unfolding method do you use?
We use TUnfold with no regularization, but cross-checked that using build-in option of Tikhonov regularization gives the same result. We also substantiated the choice by calculating condition numbers, which are small, in this case no regularization is allowed. - l185. Do we assume the extreme variations are excluded?
Yes, added this to the text.

- l217. We find strange that you use that tool
This is recommended tool for combining distributions from different channels.

- l223. How did you define that maximum value? did you apply that cut-off to the jet pt values? you should say something about it
We added formula for calculating of fiducial cross-section in Results section. There we clarify, that measured number of charm events is used.

- l223. How do you split the experimental and theoretical uncertainties? We assume the values come from Table 5, but you don't say it. In fact, it would be great to add the total ll uncertainty in Table 5, since this is what you quote in the main text. added Experimental systematic uncertaitnies include those, related to \PQc tag/mistag rates, JER, JES, identification and isolation, pileup and luminosity. The rest are theoretical systematic uncertainties.

-Table with data and fitted yields is missing. You need to add non-Z background, Z+LF, Z+b, and Z+c, both muons and electrons, this is absolutely needed We keep SFs to show postfit results for MC.

- l236. You mention 'several' generators, but you have already mentioned one in the previous sentence. Maybe you can say "other' generators. Or swap the two sentences. Summary changed: we keep it this way, because then we added some comments about all three generators after that sentence .

- l239. Do you really need the last sentence? Since you are not doing it, it's not clear why you need to mention it. If you want to keep it, we would say something like "These results are sensitive the PDF of the charm quark and can improve the existing constraints" Summary was changed, fixed.

- The summary is unclear to me, what's our take with the large data/prediction disagreement?
Summary was changed.

- Reference 15. Don't we have something better than that reference? This is usual reference for ttbar

Type-B (physics questions) + Figures + Tables:

General: Do you think a measurement of Z+charm meson production would be useful for the area of research of your paper? CMS is able to reliably reconstruct D*+, D0 and D+ mesons in pt range from as low as 3-5 GeV and up to ~100 GeV. Charm mesons would likely have smaller signal yield compared to charm jets, but, on the other hand, it is significantly easier to separate them from light and beauty contributions: it can be done by a fit to invariant mass distribution. Such an analysis could be less dependent on MC predictions. As an example of what I am talking about, you can look at this early LHCb paper done with 2011 data set: https://arxiv.org/abs/1401.3245 (we have much more data and better acceptance than this LHCb paper). (Consider citing it in the introduction). Another example is the ongoing Z/W + Upsilon production analysis BPH-19-004, where a 2D fit to m(Z)&m(Y) candidates is used to get rid of non-Z and non-Y backgrounds.

This measurement is out of the scope of this study

General: (just a curiosity) Given that you have reconstructed a large sample of Z+c events, have you looked at M(Zc) distribution, to see any indication of a new particle decaying into Z+c ? (Of course I understand that a proper search for resonances decaying into Z+c is completely different analysis, out of the scope of this paper) This measurement is out of the scope of this study General: Is only single parton scattering process (SPS) considered for production of Z+c ? Do you have any estimates on the magnitude of the double parton scattering (DPS) contribution ? Presumably it is very low due to tight requirement on c jet momentum. Maybe this could be mentioned in the introduction. Do all MC samples described in section 3 only use SPS ? Can e.g. deltaR(Z, c jet) distribution help us to distinguish between SPS and DPS contributions ?

This question requires separate study.

L25: "secondary vertex mass" was not defined, not sure if this is well-known enough term. consider adding a definition or a reference. added reference to IVF algorithm for SV, before "The invariant mass of tracks associated with the secondary..."

L59-61 In my opinion, it would help to clarify which (at least dominant) processes with top quarks and double boson production contribute to the background in your analysis. For instance for double boson, I imagine, it is ZW, where W decays into charm and strange jets. There is a short description of backgrounds: top (single and ttbar) and dibosons. Given that their contribution is small, we feel that we don't need to go into the details of how these different processes may mimic the signal event.

Such process is possible, but contribution is too small. Single top + dibosons / DY ~ 0.01 (after c-tagging)

L78: quality of the fit - which fit ? track fit ? (consider adding a clarification)

L78: please specify how is the primary vertex selected. Fixed: the fit quality -> the tracks fit quality

Fixed

L102: how are the pileup vertices defined ? All the other vertices except the "main" one ? If yes, is the main vertex selected as the one associated to Z->ll candidate ? Yes, lepton id includes cut on distance from primary vertex, which is defined as one with highest sum pt.

L112: same question. How do you ensure the charm jet originates from the same PV where Z boson is produced. Is there any requirement on dz between Z decay vertex and charm jet production vertex ? PileUp id is applied to select jets associated with primary vertex.

L128: "c tagging rate" is not clear: do you mean efficiency ? Fixed

L133: does this selection affect the reported measurement? (It is corrected for ?). For example I could imagine artificially increased cross section measured in high-pt(c) bins due to this requirement. What is the average number of c-tagged jets before this selection is applied ? Selections at reco level can only increase/decrease acceptance/background, since measurement is divided/multiplied by acceptance/background, any selections at reco level can increase/decrease measurement uncertainties. %END%

L144: suggest to define in this sentence M_SV variable that is used in the X axis title of Figure 2. For example, "... vertex (M_SV) in the ..." fixed: The invariant mass of tracks associated with the secondary vertex ($M_{SV}$)

L145-146 if the previous comment is applied, you can simplify the sentence as "... observed M_SV distributions in the muon ..."

L150-151: please provide more details on the fits. Are they binned chi2 fits ? What parameters are free and which are floating in the fits ? Presumably all the shapes are fixed to MC and only the 3 scale factors are floating. added in the beggining of the sentence : "Binned maximum likelihood template fits are performed..."

L152: It is not clear for me why the SF_q are floating in the fits, while "top and dibosons" contribution is fixed. Is there any reason to trust MC prediction for top and dibosons more than we trust MC for light, c- and b-jets ? SVM distribution doesn' allow to disciminate all components, on the other hand contribution from top and dibosons is small comparing to DY, so shape mismodelling of these components can be neglected.

L148-154: If I understand correctly, this procedure assumes there is no background under Z boson. So that all the selected Z+jet candidates have true Z bozon and some jet, which may be c, b, or light. Is it true ? Has it been checked ? In general, I think the reader would be curious to see the Z->mumu and Z-ee invariant mass plots, showing a clear Z peak on top of (presumably) very small background, in a paper that studies Z associated production. I encourage you to add such a plot(s) into the paper or supplementary material. Actually no assumption about absense of background under Z boson was made. It is stated at the beginning of the paragraph that "The top quark and diboson background predictions are taken directly from simulation." and we believe this sentence is clear about how the background is treated.

L156: I think it would be useful to underline here that after accounting for the measured SFs, all the M_SV distributions are in a good agreement with sum of MC templates, as Figure 3 shows. Fixed

L162: may be not clear what "this signal phase space" refers to, as the previous sentence describes removal of double counting. Consider moving sentence in L160-161 somewhere else, or repeating your region definition in L162. Changed to THE signal phasespace

L175-176 and Figure 5: why are the efficiencies so much different between electrons and muons ? L176 says that the efficiency loss is due to tagging. Why does the tagging efficiency depend on Z reconstruction channel ? Muons and electrons have differnet reconstruction efficiency%END%

L178: suggest starting the section with more general sentence. For example: This section describes the sources of systematic uncertainties considered for the measured Z+c jet production cross sections. Uncertainties are estimated by varying input parameters and repeating .... We prefere to keep the sentence as it is.

L180-181: this is not clear. If I understand the procedure correctly, I would suggest to write something like "Uncertainty related to each source is estimated as the difference, in each bin, between the obtained cross section and the baseline result obtained in Section 5." The current text "the differences between the unfolded distributions" may be interpreted in several ways.

L181: it is not clear which of the uncertainties below are pt(Z)-and pt(c)-dependent (and calculated for each bin), and which ones are independent on pt(Z) and pt(c). Consider adding a clarification. Table 5 summarises uncertainties for different variables/channels.

L1822-183: Maybe these are standard and well-known scales and procedures, but they are not clear to me. The procedure also needs more explanation. If possible, give a reference. Are these scales used in the MC generation parameters ? Would it make sense to introduce them in section 3? Do you produce additional MC samples with different mu_R and mu_F to calculate this uncertainty ? (Or do you somehow reqeight the existing MC samples?) Are mu_R and mu_F scale uncertainties considered uncorrelated and summed in quadrature ? Or are they correlated ? This is standard description of qcd uncertainties.

L186-188: Maybe I am mistunderstanding something, but if your measurement is based on PDFs, there may be a vicious circle: in Line 6 you say that your measurement provides information on PDFs, but (presumably) you use the predictions for these PDFs in your measurement and discuss the related systematic uncertainty in L186-188.

This problem is common to most measurements that constraint the pdfs. First of all the pdf uncertainties affects both signal and backgrounds in this measurement. For the background both the uncertainty on the total cross section and the shape of the observables can be relevant. For the signal and other Z+b and Z+uds backgrounds, that are fitted to the data, the pdf uncertainty can have an effect only on the shape of the distributions and, indirectly, to the result of the measurement. However the latter can have a smaller uncertainty than in the pdf used to extract it and in such case it provides a useful information to further constraint it.

L192: Please clarify where do you take this standard deviation from. Are those from MC ? From some reference with a study of CSV tagger performance (give a reference in this case) ? Or these are the uncertainties obtained in the fits described in the beginning of Section 5? In the latter case, uncertainties on SF_l, SF_c, and SF_b are correlated with each other, and the systematic uncertainty obtained by varying each SF by +-1sigma may be underestimated or overestimated. The magnitudes are added. The description of these uncertainties is available only within CMS, so can't be cited.

L194-198: It is not clear where the magnitudes of the tested variations are taken from (both L197 and 198). Please clarify, e.g. add a reference. The magnitudes are added. The description of these uncertainties is available only within CMS, so can't be cited.

L199-200: It is not clear how this cross section variation accounts for pileup-related uncertainty. Please explain what is uncertainty related to pileup (where exactly does pileup enter your measurement ?). Why the variation of 4.6% is taken ? (Give a reference ?). Fixed

L203-205: Again, give a reference to motivate the chosen variations. The resulted uncertainties were obtained in this measurement and can vary in different analysis

L208: add a reference for this 10% uncertainty or explain why it is a reasonable value. Why only the uncertainty in top pair cross section is calculated. Are the uncertainties in single top and dibisons negligible ? If so, clarify it in the text. Fixed

L209: Add a reference for this uncertainty in the luminosity. Consider adding a clarification that the luminosity is used in section 7 to calculate the results. Fixed

L211: Are all uncertainties in Table 5 independent on Z or c jet pt ? If yes, clarify it in the beginning of section 6, if not, clarify what Table 5 gives. It is added to the text, that the table summarizes errors for fiducial cross-section

L219: why are the luminosity uncertainties considered to be uncorrelated ? Same question for QCD renormalization scale and PDFs.

luminosity uncertainty is correlated and taken into account as change in normalization of final results after unfolding and combining(added to the text), QCD and PDFs uncertainties are calculated so that it is impossible to track separate variations of QCD renorm scale and pdf choice.

L223: how is this inclusive cross section calculated ? Please explain in the paper. What is the last uncertainty "th" ? It has not been discussed in the text/defined.

L222-223: why only Z pt<300GeV requirement is given here ? no pt(c jet)>30 GeV requirement is listed. no requirements on Z decay products are listed.It is mentioned, that fiducial cross-section is measured, so cut on Zpt is additional to fiducial volume selections.

L222-223: suggest to reword as "inclusive fiducial cross section of Z and charm jet production for .... is measured to be 413+-...., which is significantly below the prediction of MG5... generator of 524...."

L230: according to L119, pt(e)>29GeV requirement is applied. Please clarify. The cut was chosen so that is was ~2GeV higher corresponding trigger threshold.

L234: why only pt(Z)<300 cut is listed here among all the applied kinematic cuts ? It is mentioned, that this value is for fiducial cross-section (fiducial volume is defined), this cut is additional to fiducial volume definition.

L239: "fit the PDF of the charm quark" - not clear what is meant by "fit". Maybe better "to determine" or "to estimate" or "to evaluate" ? changed to "...can improve the existing constraints ont the parton ..."

Tables 1-5:
- In general, are tables 1-4 helpful to the reader ? Would not it be better for readability to present these tables as Figures ? (like you do in Figures 4-5 for acceptance and efficiency)
The plots were changed to tables, as it is common way to present flavor components SFs - table captions should be above table
Fixed - remove outer table lines and most of the inner lines, according to the guidelines.
--- E.g. for tables 1-4, leave only the horizontal line after the 1st row and three vertical lines separating the columns
- Use double hyphen in latex to indicate the range in the 1st column of tables 1-4
- use proper minus symbol (not dash) in table 5
- use triple dash in the last two cells of Table 5
- There is no point in repeating (%) in each column title of Table 5, remove those and change the table caption to "Systematic uncertainties in percent in the measured inclusive fiducial cross section."
- Tables 1-4: increase the line spacing so the numbers don't overlap.

Figures:
- all axis labels, axis titles, and, especially, legends, are too small. Please increase their size for better readability.
- Figure 2-6: suggest to reduce or eliminate space between top and bottom panels in each figure
Fixed - Figure 2-6: the plot on the right seems to stick out from the textwidth
- Figure 2-3: marker in the legend for data points does not match the plots: there should be no line in the legend
Fixed, vertical line kept: other papers have such markers in the legend.

- Figure 2-3: green box in the legend for "top and dibosons" has a small black dot in its center, which should be removed (as there is no such marker in the plot itself)
Fixed

- Figure 2-3: to be consistent with text, remove dashes in the legend (c-jets and b-jets -> c jets, b jets)
Fixed

- Figure 4,5: make sure you use CMS TDR style script to produce the plots. Axes labels and titles are too small.
- Figure 4: figure appears to be quite empty and take a lot of space. Consider reducing Y axis range.
Fixed

- Figure 5: Suggest to reduce Y axis range and put legend below points.
Fixed

- Figure 4,5,6: figures seem to be wider than textwidth.
- Figure 4,5: are vertical uncertainties zero or too small to be visible on the plots ? Please clarify in the caption.
Yes, these uncertainties are small.

- Figure 4,5: consider using different line colors for electrons and muons (in addition to different markers and marker colors)

Do we need different colors for bars? Error bars are usually black. - Figure 4: some points are overlapping between muons and electrons, so they are hardly distinguished.

If values are close, it is hard to distinguish them and their plots overlap

- Figures 3-6: p in symbol for p_T should be in italics to be consistent with text. (X and Y axis titles)

We use standard symbol for pt in plots

- Figure 6: (top panel) legend entry for "Measurement" has a horizontal line, which is not there on the points. Remove it from the legend.

- Figure 6: (top panel) are vertical uncertainties too small to be visible on the plots ? Or they are given by the height of the yellow band ? Consider clarifying in the caption. Consider also giving the results also in a table (e.g. in supplementary material), so that one can precisely know the measured value and uncertainty in each bin (e.g. for comparison to other experiment data or new theoretical predictions).
Yes, these uncertainties are small.

- Figure 6: (top panel) The blue points appear darker than the respective blue legend entry.
- Figure 6: (bottom panel) legend entry for "Stat" should have vertical uncertainty line.
They have the same color

- Figure 6: (bottom panel) legend: Stat -> Stat. uncertainty
- Fixed - removed

Figure 6: (bottom panel) legend: Total -> Total uncertainty

Fixed - removed

- Figure 6: (bottom panel) legend: Th Syst -> Theor. syst. uncertainty
Fixed - Figure 6: (bottom panel) legend: Exp Syst -> Exp. syst. uncertainty Fixed

General Comments

There is not much mention in the paper on how it compares to the earlier 8 TeV analysis in ref 4, eg, on improved quality/improved understanding of the data/process. We should comment on that in the
discussion.
- by how much is the precision of the data (stat error) improved?

Comparing with results in SMP-15-009: total xs x branching 8.8±0.5(stat)±0.6(syst)pb, or 5.5% stat and 6.8% syst at 8 TeV, with current result 1.9% stat and 5.5% syst.

- Madgraph LO and NLO seemed to agree with the data in ref[4], now the NLO one is
20% too high. MCFM at the time was about a factor two too low (we do not test it here)
I s the all consistent with the present results? The NLO prediction is off in this paper
- Do our measured values in the two papers agree (modulo an expected CM energy dependence)?
- Given the discrepancies we see with data and calculations, I am not sure we actually
understand exactly what we can concluding. Do we consider these measurements + the theory
to describe them satisfactory and understood enough to suggest to use these data for PDF charm
extraction/constraints?
We conclude the paper saying that these results can help to fit the PDFs, but we also say before the
best agreement is reached with a LO program QCD program. The way it stands there now, it looks
like a contradiction to me. LO predicts less jets comparing to NLO, but the prediction for charm component may be lower, than they actually are, just like with NLO, so these two effects compensate each other for LO.

Details
- line 53: matching: the matching is again discussed for all MC programs on line 66. Is that not
sufficient?
FIXED

- line 54: So the madgraph (LO) cross sections are not used but we use the FEWZ
ones, correct? This is used to for the curves in the comparisons in Fig 6 as well I presume?
This may not be clear when just looking at fig 6 where it looks as these
are madgraoh predictions from he legend. As people may just
see this figure in talks, one should explain this better in the legend (or at the
minimum in the caption). Similarly for ShERPA We mention that all 3 signal models are normalized to nnlo with fewz in section, where MC models are described

- line 65: is this 19 GeV the default value for MLM? Or why was this value chosen and
not say, a round number like 20 GeV?
There is a standard procedure to choose the matching scale for a given process. Even if not so important, it is a useful number and should be quoted for completeness. There is no need to explain how this is chosen, which is a rather technical detail, well known to generator experts.

- line 69: Geant4 is written in the wrong format, for what we usually do in CMS papers.
ps: we use: \new-command{\GEANTfour} {{\textsc{Geant4}}\xspace}
FIxed

- line 76: add a ref to the muon reconstruction paper (in order to make it symmetric with
what we do now for the electron paragraph following this one).
Added the same reference as for muon pt resolution, the same paper describes muon measurement parameters at CMS

- line 78: Say somewhere in this section how the primary vertex is defined in this
analysis
Fixed.

- line 86: How do we compensate for pile-up contribution? Add a reference for the method used.
I can't find reference to this part in other papers, it seems there are some technical details, which are available only on some special POG twiki.

- line 92: are the electrons required to originate from the primary vertex with the
same conditions as above imposed on the the muons? please specify what is used
Fixed. - line 98: [r1][r2] -> [r1,r2]

- line 125: since c-jets are very central to the physics measurement in this paper
I would have expected a few more lines on explaining how the c-tagger that we use
in this analysis works, not just a reference. Some information, that it uses different parameters of the jet, such as SV or its impact parameter, is added.

- line 133: So we don't take the c-quark with the highest quality as given by the
c-tagger? Would that not be the better choice? I assume you checked that?
Not sure what highest quality means here, the tagger shows, whether jet passes the criteria or not. Since we select highest pt c-jet at gen level, there is more change to match with it if we select highest pt c-tagged jet.

- line 170: what used as standard MC for the response matrix for the unfolding?
What systematic uncertainty is taken for the unfolding, eg the response of alternative
MC models? There is no systematics in table 5 for the unfolding.
Please spell out what is done.
We did unfolding of measured Z+cjet events using acceptance, responcse and background from LO madgraph, results are in agrrement within stat errors with results from unfolding, using the same variables from NLO madgraph, so unfolding with LO madgraph wasn't used as uncertainty

- line 185: The typical question on how the scale uncertainty was determined.
Do we avoid in this process to include combinations where the change of the ratio of
the scales is more than 2? It is strongly recommenced NOT to use these extreme
variations as otherwise you get a too pessimistic systematic uncertainty from that
source.
Yes, unticorrelated combination 2muF,0.5muR and 0.5muF, 2muR are not taken into account.

line 187 what is "CT14 [59]" There are only 39 references. Is that something
left over in the text?
This is an error, this pdf set shouldn't have been mentioned, fixed.

- line 188: You use the NNPDF PDF distributions so why you use here the
CT14 description for the uncertainty and not the NNPDF one (or the PDF4LHC one)?
- line 200 one could add a reference here for this number (as we do in other papers)
This is an error, this pdf set shouldn't have been mentioned, fixed.

- Table 5: +/- uncertainties for one channel/quantity are not always given to the same
numerical precision. This looks sloppy. fixed.

- line 217-224: The discussion at the end of this paragraph is very terse. See also my general
comments above. Is it not weird that the NLO calculation is so significantly off?
Or is it because of the FEWZ normalization that the LO one works better?

It might be more complicated: Z+ any jets NLO provide better agreement with data, then LO, normalized to NNLO, but overshoot Z+c-jet. LO might overestimate Z+c-jet, but since total Z+anyjet is underestimated, this effect is compensated and LO provides better agreement.

Or can we blame the charm PDFs (probably not). The reader is a bit left on his hunger here.
- line 237: ... but this is with NNLO cross sections right? So this is a bit of
an unfair comment in that case. At least this needs to be explained. Added, that predictionas are normalized to next-to-next-to-leading order cross-section

- l.150 could you specify what you mean by fitting.
Is this a Poisson based maximum likelihood template fit or
is it a chi2 fit?

Fitter was changed to TFractionFItter which uses likelihood fit.

- l.150 You are not discussing the problem of limited MC statistics
for the Z+c, Z+b and Z+l tempates that are fitted to the data Msv distributions.
Can it be safely neglected? Normally one uses the Barlow Beeston method that
is implemented in the TFractionFitter code. Or one uses the combine tool
where the so-called Barlow Beeston lite method is used, approximating
the uncertainty of the sum of templates by a gaussian distribution.
In particular for the fits in the differential bins one could easily
imagine that the MC statistics could play a role in the fits.

Yes, RooFit doesn't take into account stat errors of MC, analysis was redone using TFractionFitter.

- Fig. 2 as discussed in recent months the SV mass is alreay sculptured/distorted
from the deep CSV tight c tagging criteria that already used
the mass in order to enhance the c component. I think this distortion
should be mentioned in the text. Otherwise it looks like we have
much worse SV separation for the various flavours than other experiments.
Normally the b-jets are the only source that have a large contribution
between 2.5 GeV and 6 GeV, but here, after the deep CSV cut, the situation
has changed. We can still say that after all we have still a reasonable
separation power. In this content I think it would be very much worthwhile
to also mention in the paper the cross check you performed with using
the deep CVS b medium point, after which the SV mass looks much different
and less sculptured. If I recall correctly the scale factor results
(table 1 and 2) come out consistently.

Yes, we have checked, that after using medium b-tagger instead of tight c-tagger the unfolding results in both cases are consistent withing errors (stat and tag SFs unc were taken into consideration), not sure, what is the best way to put it in the paper We added the sentence, that MSV is used in c-tagger, though still can descriminate flavor components. We also decided not to add results of cross-check using b-tagger.

- Fig. 2 it would be really nice to show how the post-fit SV mass distribution looks like, so how the template fit really describes the SV mass discriminator spectrum after the fit. One could do that using the total sample like in Fig. 2 or after summing up all the fits in five bins of Z candidate pT, separately for electron and muon channel. Such a Figure is missing in the paper. One could also see better if some residual shape problems that are visible in Figure 2 vanish with the fit adjusted templates. After all you have not really demonstrated that data are described well by the fit. (One step further would be some chi2 GOF test). It is nice to see the control plots in Figure 3 but still a post-fit plot of M_SV would be good to see. Final purpose of the fit is not good agreement of data/MC for SVM, but for pt spectra, do we really need to include these plots? MSV distributions after applying k-factors as functions of Z or c-tagged jet pt are added as suplementary material.

- Table 1 and 2: I would add some information about the level of fit-correlation
between the sources, as usually expressed by the correlation coefficients rho.
State some typical value for rho_bc, rho_clf and rho_blf in the text.
I would guess that rho_bc comes out to be negative and somewhere between
-0.5 and -0.9. Of course if it would be -0.99 it would become
pathological.

The correlation coefficients between charm and bottom components are shown in plots in attachment (correlations.pdf) as function of pt. As expected by Olaf, these are ~-0.75-0.85.

- l.169-172: the unfolding is missing the information
that you unfold the fitted Z+c event numbers (which you get
from the product of SFc times the number of Z+c signal MC events
at detector level, per bin). Please add such a sentence.
Are you using really the same number of bins at detector level
and generator level, or you are using more bins at detector
level, which is what TUnfold normally expects. We have been
probably already discussing this recently but I forgot what you did.
If you use more bins at detector level I think this should
be clearly mentioned in the text. If you use more bins it
is a generalised matrix inversion (from chi2 minimization), matrix inversion
would be used for the same number of bins at both level. Added phrase tha ttunfold is used to unfold measured Z+cjet events distribution, and Nrecp = 5 and Ngen =4.

l.127 and l.189-193 the information on the flavour tagging calibration
and its uncertainty is too brief in my view, or there should be
some references. It seems, that information/plots about this SFs calculation is available only within CMS, thus we decided to leave it as it is.

Fig. 3 is there a problem in pT^ll in the first bin? It looks like the residual mismodelling of Z pt spectra: in case, when no c-tagger is applied the "wave" in data and MC ratio is observed for small Z pt values.

l.217 "The results" which results??? The cross section results, right? Fixed

l.224 How is the theory uncertainty 11.7 (th) pb calculated?
should be described in the text. Fixed

l.223 how was the inclusive fiducial Z+c cross-section measured, is it
just from summing the differential cross sections vs pT^Z?
I suppose so, but where is that stated in the paper? Added formula to the results section

l.223-224: there is really a significant discrepancy betwen data and a NLO
calculation, isn't the large level of discrepancy a bit surprising
and we should think further about it?
In this context: do I understand correctly the for
the signal MC NNPDF3.1 PDF set is used, at NLO,
and in the 5 flavour scheme, including c and b quarks
in the proton PDFs? Does the simulation also include processes with an inital gluon in one of the proton that splits into a ccbar pair and one the charm quarks radiates a Z and interacts with a gluon from the other proton? Yes, c-jets can originate from gluon splitting, but this measurement is inclusive Z+cjet and takes into account such jets

PHYSICS COMMENTS:

• Title: ... and charm jets in proton-proton collisions at s√=13 TeV. Fixed

• - LL7-12: please split this very long sentence in two - it's hardly parsable otherwise. Fixed

• - L40: within a fixed time interval of about 4 μs. Fixed

• - L47: ... \MGvATNLO 2.2.2 [8,9] (MG5\_aMC) ....

• - L48: please, specify the perturbative order at which the Z+c jet process is simulated. Is it NLO, just as for the Z+jet backgrounds? Fixed

• - L54: give the full version of FEWZ here - is it 3.0? Fixed

• - L55: ... (n=0--4) processes. The value ... [no need to specify the MLM matching here, as it is already specified on L85, which should be moved before the present sentence [see a comment on LL2-67 below]. Also, give the full SHERPA version here; is it 1.0? Fixed

• - L60: give the full \POWHEG version here, 2.0. Fixed

• - LL62-67: it makes little sense to discuss the matching schemes until you describe the parton shower generator. Therefore, it would be logical to move this paragraph right after the first sentence of the section and then start "The \MGvATNLO version 2.2.2 ..." as a new paragraph. Also, is this really true that you use PYTHIA to shower the SHERPA samples, given that SHERPA has it's own parton showering and fragmentation built in? Also, please give the full PYTHIA version here, 8.212? On LL62-63, say: "... showers, hadronization, and the underlying event, with the CUETP8M1 tune [21] that uses the NNPDF2.3 [22] LO PDFs ...".

Fixed: madgraph and Sherpa use nnpdf3.0, then pythia uses cuetp8m1 or cuetp8m2t4 with Nnpdf 2.3

• - L78: define the primary vertex here. Fixed

• - L86: either give a reference to, or explain how the isolation is corrected for the effects of pileup here.Fixed

• - LL76-96: you constantly switch the order you describe electrons and muons in the paper, which is pretty annoying. Please, use the same order - electrons before muons - throughout the paper. Here you need to swap the order of the two paragraphs. Now, the muons are required to be isolated, but you say nothing about the electron isolation. Surely, you require electrons to be isolated too, so please specify the isolation criteria in the corresponding paragraph. Similarly, you talk about the dielectron mass resolution, but muon pT resolution. This makes little sense, so please either specify the dilepton mass resolution for both leptons, as this is the relevant quantity for this paper.Fixed

• - LL113-114: again, swap the order of the electron and muon trigger description here. Fixed

• - LL115-119: ... and the order of the electron and muon paragraphs here. Fixed

• - Tables 1-4: swan the order of Tables 1,2 and 3,4 [i.e., 2,1,4,3], so that you talk about electrons before muons. Also, suggest identifying which pT is used in each table fro clarity [i.e., use pjT in the headings of Tables 1-2 and pZT in the headings of Tables 3-4. Fixed

• - LL158-159: I don't understand why the generator-level phase space is defined differently from the reconstructed one. In the reconstruction, you use pT>29 (26) GeV for the leading lepton in the electron (muon) channel, while the generator-level phase space is defined as pT>26 GeV. You can't be that sloppy in a precision measurement paper, particularly since you claim that the fiducial cross section you measure is significantly higher than the theoretical one within the same phase space! Surely you will measure lower cross section in the electron channel given a harder lepton pT requirement! I therefore believe that the measurement should be redone in the phase space defined by the leading lepton with pT>29 GeV, both at the reconstructed and generator levels; only then the comparison would make sense. The phase spaces at gen level and reco level don't have to be the same: the difference is taken into account by fraction - acceptance. Maximum of distribution of leading electron pt is bigger, than 29 GeV, so the gap doesn't have big impact on acceptance, which is mostly contributed by c-tagging and electron id efficiency. On the other hand, muon sample has larger statistics, so the threshold was taken as low as possible.

• - Figure 2: swap the order of the left and right panes and update the caption accordingly. It would really be useful to also show the same distributions after the scale factors are applied, either as a second set of plots, or as the second Data/MC distribution in each lower pane of the present plot. In the lower pane y axis legends, use "Data/MC". In the legends, use: "LF jets", "c jets", "b jets", "t quark and dibosons".

Fixed

• - L172: specify what unfolding method was used - d'Agostini, SVD, ...? Fixed

• - L182-183: Renormalization and factorization scales: The ambiguity in the choice of the renormalization (μR) and factorization (μF) scales leads to an uncertainty ... [These are not QCD sales! While the renormalization scale, is indeed a scale of a QCD renormalization group equation evolution, the factorization scale has to do with PDFs, and not QCD matrix elements.] Not sure, it is not stated, that these are qcd scales.

• - L185: specify the default μR=μF scale used in the calculations. Fixed

• - LL186-188: first of all, Ref. [59] is nowhere to be found; the paper only has 39 references, none of which resembling the CT10 one. Second, why do you use a CT10 uncertainty set, rather than the full PDF4LHC prescription for Run 2?

The whole paragraph seems wrong: changed to smth more common, like in other papers.

• - Figure 2: swap the order of the two rows of plots - electrons before muons - and update the caption accordingly Fixed. In the lower pane y axis legends, use "Data/MC". In the legends, use: "LF jets", "c jets", "b jets", "t quark and dibosons". Move the x axis labels a bit lower, so that they do not overlap with the axis values.

• - L200: give a reference to our total inelastic cross section measurement to justify the 4.6\% number. Fixed

• - L208: why the cross section is varied by 10\% - where does this number come from? Please, provide a reference. Fixed

• - L210: give a reference to LUM-13-001 for the 2.5\% number. Fixed CMS-PAS-LUM-17-001

• - Figures 4-5: swap the order of the electron and muon keys in the legend.Fixed

• - Table 5: the caption states that the uncertainties are in the inclusive cross sections, whereas the entries clearly show differential results, as the uncertainty depends on whether the pT of the c jet or the Z boson was used. Please correct the caption accordingly. Changed "inclusive" to "integral".

• - L223: explain "(exp)" as experimental systematic uncertainties. Fixed

• - L224: how was the theoretical uncertainty estimated - please explain in the Systematic uncertainties section. Does in include the scales and the PDF? Fixed

• - Figure 6: the legend is illegibly small - please make it bigger. Move the x axis labels lower so that they do not overlap with the axis values. In the lower pane legends use "Theo Syst.".

Fixed

draft V6 comments

draft V4 comments

ARC questions

Philip

Luca

Cecile

Vitaliano

Answers (CMS Statistics Questionnaire)

Answers (pre-approval)

Answers (paper draft v1) Juan Pablo

Elisabetta

Answers (paper draft v0) Elisabetta

Joel

Answers (paper draft) Elisabetta

Juan Pablo

Answers (AN) Elisabetta

Juan Pablo

Isabel

Paolo