This is one of the important papers for historical oceanography and it focuses in the reconstruction of environmental measurements translating them in modern units from literature records. I recommend the publication subject to minor revisions mainly adding better images, more discussion and interpretation of the results, in addition to some nomenclature issues which I suggest should be corrected by the authors before publication.

I have some major remarks:

• I believe it would be nice to have the Marsili’s maps reproduced ( Gulf of Lion map - Pl. I, page 3 and Cassis map - Pl. II, page 4) because the discussion is otherwise very difficult to follow. There is somehow a little confusion between “distance” along the transects and “lengths” discussed. I suggest that instead of “lengths”, “distance along transects” is used.
• Section 4.1 should not be called “Bathymetry: historical map analysis” but “Measurement sections: historical map analysis“. Furthermore several times there is “map” instead of “distance”. In oceanography we do not use the word map to indicate distance on the earth.
• Furthermore why calculate the “mean” distance? The text is reported here: [The mean distances are 2912 ± 6057 m for the Gulf du Lion map and 324 ± 472m for the Cassis map.] What is the oceanographic significance of a mean “section distance”? I suggest to eliminate this.
• Furthermore I believe the authors should conclude in a more precise manner that to have distances matching the present ones you need to consider a different coastline, as shown in Appendix A. This is an important consideration which is not presented adequately. 
• Your statement: The minimum possible measurement error has been calculated, yielding results of ±1.23 Kg/m³. Please define how you estimated errors, this is important. Furthermore, why showing Fig. 3 if the amphores were shown already in Fig. 1?
• Throughout the text, even in the conclusions, it is not mentioned that Marsili did his first measurements of density in the Bosphorus, much before the ones in the Gulf of Lion. In addition he used a different reference water, this is in my opinion an important methodological aspect. Please insert somewhere this relevant note and refer to Pinardi et al. (2018) for the discussion on the other reference waters.

Detailed comments

• Please give a DOI for the EMODnet_satellite_coastline_MSL, it cannot be referenced in this way;
• Line 205, the formula was first presented by Pinardi et al. (2018) and it should be referenced;
• From line 270-280: I suggest you do not talk about the history of past projects/efforts to collect the data in the SeaDataNet archive but you just reference the database which has a DOI. This is not a paper on the recent historical database and it should be enough to give its referencing in modern literature.

Everywhere:

Lenght should be changed to length