Gola et al. (2005) and Sellevåg et al. (2006) have neglected the influence of the reaction of CH3Cl with O(1D) in their experiments. It is also missing from their FACSIMILE-based kinetic reaction modelling (supplement to the Gola et al. paper). This reaction is likely to have biased their results for the isotope effect of the CH3Cl + OH sink. In the present set of experiments, the influence of the O(1D) reaction is likely to be small (assuming the difference between the measured CH4 + OH isotopic fractionation and the literature value of Saueressig et al. 2001 is due to the presence of O(1D)). See also comments below. You should make a stronger statement (around 7/24; "unresolved" doesn't really capture the potential experimental error of the Gola et al. experiments).
It would be interesting (and straightforward) to measure the isotope effect of the reaction of O(1D) with CH3Cl. You do not appear to have done a corresponding control experiment (O3 photolysis without water vapour). Please mention the absence of this control experiment in your discussion.
Use decimal points, not commas, throughout the manuscript (incl. figures and tables).
You sometime refer to control and sometimes to "blank" experiments. Please use the term control experiment throughout the paper.
1/22: For clarity, please write "13C/12C carbon isotope effect (ε = k(13C)/k(12C) – 1)" (or "carbon isotope fractionation").
1/23 & 1/24: Please write either "we found an ε value of (x±y) ‰" or "we found ε = (x±y) ‰."
1/27: Replace "fractionation factors" with "values for the carbon isotope effect" or "carbon isotope fractionations".
2/18: Change to "KIEs of the main tropospheric sink reactions (CH3Cl + OH, CH3Cl + Cl)".
3/37: Brackets missign around (-26.8±0.2) ‰.
4/1: Measurements against the machine working as do not follow the "identical treatment principle", so the value of –37.2 ‰ should be provided for anecdotal purposes only. This is NOT an offset between DI and GC-IRMS method. The offline method represents the calibration of the CH3Cl standard, which should be used as anchor point to report all subsequent measurements of CH3Cl by GC-IRMS.
How often was the CH3Cl/N2 reference standard analysed?
Was the same CH3Cl/N2 reference standard used in Hamburg and Heidelberg?
4/15 & elsewhere: Please change to "the kinetic isotope effect (KIE, symbol ε)". To avoid ambiguity, whenever you refer to the value of the KIE in the manuscript, please use the quantity symbol ε rather than the abbreviation KIE, including in the supplement.
4/15 & 19 (3 times): Delete the square brackets around ‰.
4/17: Replace "1000" with "1" (two occasions).
4/18: Replace "enrichment factor" with "kinetic isotope fractionation" or "kinetic isotope effect".
Change to "residual CH3Cl fraction".
5/4: Negative sign missing before "0.00117".
5/12: Delete "molecule–1". There is no such unit.
5/12: Replace "*" multiplication signs with "×".
5/16: Please change to "the partial lifetime of CH3Cl with respect to OH" (the OH lifetime is much shorter).
5/20: "stable carbon isotope delta values"
5/20 to end of page: Replace "KIEs" with "isotope effects" or "isotopic fractionations" or "ε = …", as appropriate.
5/30: Insert "(in terms of absolute magnitude)" after "upper end".
5/37: A 3 % contribution from O(1D)-related loss would have only changed the epsilon value by about 0.3 ‰. This cannot explain the difference to the result reported by Saueressig et al.
In fact, a contribution of 9 % (= (4.7–3.9) / (13–3.9)) from O(1D)-related loss is required to explain the higher (in terms of magnitude) ε value found here.
7/1: Based on the value of 9 %, the potential contribution of O(1D)-related loss to the overall CH3Cl loss in your experiment is 2.3 % - still small, but not <1 % as you claim. It is less than for CH4 because the relative reaction rate coefficients favour the reaction with OH over O(1D) in case of CH3Cl compared with CH4.
8/15: Komatsu is misspelled.
8/26: Please give the "other" NH and SH source strengths also in Gg a–1, for consistency and ease of comparison.
9/19: Without an estimate of the degree of CH3Cl break-down (e.g. soil mole fraction / atmospheric mole fraction), nothing can be said about the magnitude of the isotope effect based on the soil δ(13C) value alone. E.g. if only 5 % got broken down, the 2.4 ‰ difference between soil and atmospheric 13C/12C ratio would correspond to an kinetic isotope effect of –48.5 ‰.
10/9: Your statement "In this case, increasing the soil sink could even lead to a depletion in the tropospheric δ13C once the apparent KIE of the soil sink becomes smaller than the KIE of the OH sink." does not make sense because the tropospheric δ13C is "fixed" (by observations). Presumably, you want to say that "increasing the soil sink" leads to a decrease in the overall sink isotope effect, which must be balanced by an overall increase (i.e. becoming less negative) in the average source δ13C value.
10/23 & 10/34: The atmospheric measurements impose a strong constraint on the model and should be discussed in their own section (new section number 3.2), immediately following the model description (section 3.1) and before section 3.1.1.
Sections 3.1.1 and 3.1.2 should be renumbered 3.3 and 3.4.
Section 3.2 and 3.3 should be renumbered 3.5 and 3.6.

Supplement
Please provide consecutive line numbers for the supplement.
Some of the information in the supplement duplicates the methods section. Please remove anything from the supplement that is already included in the methods section.
Page 1
• Teflon FEP (not Felon FEP)
• Replace "ppb" with "nmol mol–1".
• <1 nmol mol–1 O3 (see page 6)
• <500 pmol mol–1 NOx (or <0.5 nmol mol–1, which helps avoid using another unit)
• "for at least 4 μmol mol-1 8 h" does not make sense
• "typically cleaned for 6 to 8 h" (delete h after 6)
• What was the make and model of the Teflon fan?
• Teflon is written with uppercase initial T because it is a proper name.
• O2 and NOx should be written with subscript "2" and "x". "x" in NOx should be written in italics.
• Delete "ppm" in "1ppm".
• Should this be "between and 1 and 10 µmol mol–1"?
• Delete "~" sign before 25 µmol mol–1 and replaced with an actual estimate of the uncertainty of this quantity.
• helium is written with lower-case "h" (or use the chemical symbol "He")
• Valco – capital V

Page 2
• (also 3/23) Please state the actual actinic flux (in photons per area and time or energy per area and time) as well as the make and model of the solar simulator, in addition to "actinic flux comparable to the sun in mid-summer in Germany".
• Please state the temperature at which the experiments were carried out.
• Delete ")" after 5 %.
• Replace "*" multiplication signs with "×".
• 25000 mmol mol–1 should be 25000 µmol mol–1. However, even that corresponds to a temperature of 27 ºC. Is this correct?
• Delete "molecule–1". There is no such unit.
• Delete "molecules" after 2.9 × 109 and after 2.0 × 1010.

Page 3:
• H2
• Replace "μmole mole-1" with "μmol mol–1" (2 times).

Page 4
• "an unknown source" (not: emissions)
• "were randomly" varied? What do you mean by random? What underlying probability density function did you use? Gaussian? Uniform? Triangular? Beta? Please be more specific.
• How many model runs did you carry out?

Page 6:
• Replace ppb with nmol mol–1 and ppt with pmol mol–1.

Figure S2: Please add a figure legend to identify the three different colours/symbols (or label the curves as in Fig. S5 below). Use decimal points, not commas. The axis labels should be y(CH3Cl)/(µmol mol–1) and y(PFH)/(µmol mol–1), to designate that you have plotted mole fractions (the square bracket symbol means "concentration").
The right axis should be labelled "residual fraction ft" (see Eq. 3; you might want to add a cross-reference to this equation in the figure caption). Use the same scale for both mole fractions (the scaling by 0.5 doesn't seem to be correct anyway based on the information on p. 3/8

The x-axis should be labelled with integer numbers (0, 2, 4, …) to avoid confusion with clock time. The axis label should be "time t/h" [this means time divided by hours].

Figure S3: Again, a figure legend is needed and the axis labels should be amended as per the comments on Fig. S2 above. The equation in the caption should be written as y(CH3Cl)/(µmol mol–1) = 133.5 e–0.004t/h. Please provide another decimal for the coefficient in the exponent.

Fig. S4: The axis labels should be amended as per the comments on Fig. S2 above.
The equations should be written as
y(CO2 measured)/(µmol mol–1) = 469.5 e–0.00118t/h and
y(CH4 measured)/(µmol mol–1) = 3.52 e–0.00117t/h
to be dimensionally correct.

Fig. S5: Again, axis labels should be "time t/h", "y(CH4)/(µmol mol–1)", "y(CO2)/(µmol mol–1)" and "y(O3)/(µmol mol–1).

Many thanks for the quick turnaround of the revised manuscript, which
addresses all of my comments very comprehensively. The paper should be of significant interest to a wide cross-section of the atmospheric science, biogeochemistry and paleo-climate communities, and I will suggest to make it a highlight paper.

It would therefore be great to make the paper as accessible to the general reader as possible, and for that the suggested earlier discussion of the atmospheric observations (and a potential renumbering of the paper’s sections) would be of great help.

I also note that you call upon these observations at the end of
section 2.3, but they do not appear until much later (section 3.2).
This would mean reordering the figures as well (Fig. 4->2, 2->3,
3->4).

Best regards
Jan Kaiser
Editor ACP