The manuscript by Hulin et al compares C inputs of 12 cover crops with a labelling approach. SOC new as a pool consisting of fine root C and rhizodeposition is determined. With this the paper provides new data and complements the existing literature on C inputs. The scientific methods are valid and reported nicely. The analysis and conclusions are sound and clear. However, the presentation of the results in the figures was not that clear (see comments below). I recommend publication after incorporating these comments.

L35: C instead of Carbon

L37f: what processes? How much does it vary over time?

L43: what did Huang et al find? Change ‘not confirmed’ to contradicted?

L44: sentence seems off

L54: reference?

L55 remove C?

L132: how many plants were grown in one mesocosm?

L148: SOC

L234: unclear. What is compared to a greenhouse?

239: hard to understand to what the literature refers, maybe you can report it like the barley

Fig 1: upper panel letters are not in the centre of the bar as in the lower panel, why no statistics for the middle panel? Statistics for lower panel confusing. Please consider displaying the SOCnew:Belowground C inputs ratio to match the statistics.

Fig 2: Visually irritating. It looks like the whole 0-20cm horizon is missing with statistics and error bars. The orange and blue “background bars” don’t add any information. I assume that the horizon 0-20cm is the percentage that is not framed and striped. Please add a proper bar for this horizon.

L308: was it expected that the Poaceae had the highest rhizosheath values? Report the mean value in the text so the reader can directly compare it to the mean across species.

L315: “our results” and then you cite a different paper. Confusing in the way it is done now.

Fig 3: Confusing with the big difference between axis and hard to read. Consider to plot it separately in two panels. Report statistics for the linear correlation in the plot and not only in the text. The inset is very small.

L322: C loss, please check for consistency throughout the manuscript

L326: reference?

L344: fresh OC = SOCnew right? Stick with one term.

L373: sentence is off

L375: the conclusions of the review article you cited are not valid without restrictions: https://doi.org/10.1111/ejss.70077

Figure S5: consider making this figure part of the manuscript instead of only using the inset in Fig 3.

This study is a convincing original study using sophisticated 13C plant labeling in order to quantify C input to soils via rhizodeposition from different plant species. The study thus addresses a relevant question within the scope of Biogeosciences.  The study shows that it is an important C input which affects soils beyond the rhizosphere. The study is conducted well, even though there are analytical limitations due to the small proportion of rhizodeposition as compared to total soil carbon, which is taken into account in the study and discussed. A key issue is the different state of plant development (phenology) among the 12 plant species and also within plant replicates which makes comparisons between the plants difficult. Maybe plant groups need to be formed with comparable developments. Some more issues listed below should be taken into account to further improve the study. In general, the paper is written well, figures and tables are nice and meaningful and conclusions are substantiated by the presented results. Thus, this study will be an important contribution to a topic that is understudied.

1. Title: Why is “cover crop” in the title but mentioned only once in the total manuscript? Alfalfa is no cover crop and also barley, rye, oat and oil seed rape are rather used as main crop and not as cover crops. On the other hand: Vetch, white mustard and camelina are typical cover crops species and I do not know them as recently grown main crop. “Plant species” would be a more appropriate term.
2. l.26-28: There maybe definition problems with the term “rhizodeposition” if some authors do not include root debris and root hairs under this term. However, rhizodeposition is defined mostly as “the total carbon transfer from plant roots to soil” (see Nguyen, 2003). Thus, this sentence distracts from the key problems. Its not a definition problem but an analytical problem. I recommend to remove the sentence.
3. l.42: “other forbs” as compared to legumes that are also forbs.
4. l.49: Citation is missing for the “promising recent works” you referred to.
5. l.82-86: This is a method section in the introduction. I recommend to shorten it or remove it, since it distracts from getting to the objectives of this study.
6. l.87: Disentangling the spatial distribution of what?
7. l. 88: “assessing SOCnew persistence…after harvest of the above ground biomass through…”
8. l.91: Please delete “simulation”
9. Throughout the manuscript I learned: Numbers up to twelve should be in letters not in numbers and numbers and do not use numbers or abbreviations at the start of a sentence.
10. l.93: “ Here, we…”
11. l.100: Use “deionized water” instead of “osmosis water”.
12. l.102: A water content of 0.5 cm3/cm3 means saturated water conditions in most soils (total pore volume of 50% is water filled). Please check and revise. Please also write how you controlled soil moisture. How regular? Using which method? How did you adjust soil moisture? Individual for each plant species or mesocosm?
13. l.104: “following a crop harvest”
14. l.118: What is “a realistic plant density”?
15. l.125: “Replicated three times across three climate chambers…”
16. l.131: How much soil was collected as composite sample out of the 13 liters soil?
17. l.148: The net rhizodeposition that you measured was accumulated during the 56 days of plant growth, likely with much more input during the last days of the experiment when the roots were well developed. There might me also some rhizodeposits that were recent (some hours before the experiment was harvested). Are those also net rhizodeposition? Some parts of rhizodeposition is already decomposed and mineralized and thus not captured. How did you account for this flux of rhizodeposition?
18. l.206: There was large variability in the number of plants per mesocosm and plant species but also among the plant species. How well does the density fits to density of crops when cultivated at field sites? How does plant density affect the results? Is plant density a major predictor for root biomass and maybe also allocation (R:S rations and SOC new)?
19. l.225: It is not helpful that results and discussion are merges into one section. This dilutes the results of this nice study and makes it more difficult to read. I strongly recommend to rewrite this section and disentangle results and discussion.
20. l.235-238: Not clear if you refer to above and below ground biomass with the term “biomass”.
21. l.250: Need to be better explained what is labelling heterogeneity? Is it a spatial heterogeneity or variability between replicates or is it heterogeneity in the label within the plants due to the pulse labelling?
22. l.275: Are there differences in the SOCnew quality (exudates vs mucilage, vs root fragments; differences in compounds and CN ratio) among the different plant species? Thus, how could the gross vs net SOCnew ration vary among plant species? Maybe you could add some discussion on these questions here.
23. l.291: Daikon radish is considered as plant for deep soil melioration (to break up soil compaction) due to the deep reaching tap root. You found the opposite. Please discuss.
24. Figure 2 is not easy to read with the mix of colors and shading. The orange and blue bars going up to 100% contain no information and could be removed.
25. l.316: This result maybe different in topsoil and subsoil (any data on this to show and discuss?). Please discuss if also subsoils (also undisturbed once outside pot experiments) receive SOCnew across the whole soil matrix as you found it. It is in any case an important finding of your paper an may be better considered in the abstract (l. 10): Which volume is rhizosheats compared to bulk soils and which fraction of SOCnew is in the rhizosheats and which is outside of them?
26. l.368: Above and in the abstract you claim SOCnew being more persistent than roots. Here you write about similar persistence. Please rewrite.
27. l.397: This “first step” is common practice in modelling studies of soil C dynamics that need to estimate C inputs including rhizodeposition (e.g. https://doi.org/10.1371/journal.pone.0256219 or https://doi.org/10.1007/s11104-022-05438-w). Pausch and Kuzyakow 2018 generated a valuable estimate for this rhizodeposition that was confiormed in your study. Thus, your study rather supports this common practice than facilitating it.