This paper is a nice comparison of gases used for DART analyses, with the focus on the potential to move away from the use of He for environmental reasons. This is well received based on the continuing issues with the environment, but also the short supply of He.

There are a couple of points that I feel need addressing:

• The introduction feels like a Class 101 to He Dart and the respected ionization mechanisms. I think this is useful for the reader, but considering the purpose of the paper is to compare He/N2/Ar, I feel that the latter two are underrepresented in this opening statement.

• It is surprising to see that the authors have based their discussion on reproducibility and (sort of) quantitative approaches, yet they have decided to remove outliers which are over 2 STD outside of the fit line. Their argument is that DART is a manual process and so is subject to wide variability of loading content, which is a perfectly valid point to the major limitation of this work for quantitative analysis. However, the paper later goes on to say that 'repeatable analysis' of molecules can be carried out, despite the authors themselves identifying that RSDs were, in some cases, over 100%. I think to truly reflect the data being generated that all analyses should be included in these data investigations.

• Again, as the authors are arguing the (semi) quantitative value of their analysis, and using linear best fit/LOD calculations to support this, I think it is imperative that the authors also provide some additional basic data for assay quantitative validation. I would suggest that authors (in the supplementary file) provide the RSDs and back-calculated accuracy values for all molecules across all calibrator levels. This will then provide a true understanding of the nature of quantitation afforded by these approaches. This is not to say the paper is not valid if these data are not strong in terms of validation, but it will allow the readers to make a more informed decision on the work.

Other minor points:

• Can the authors provide some reasoning as to the choice of amino acid analyses as they feel a little of place. The authors mentioned it was based on 'screening' applications.

• There is repeated use of casual language throughout the manuscript (e.g. better, worse, best, etc.). It would be advised to use more objective explanations for these points that use data/results to provide the explanation/reasoning.

• Page 7 Line 15 - Should read "This suggests"

• Data are plural (e.g. page 10 line 16).