The paper describes a meteorological record covering the late 17th and early 18th century for Paris. Given the length of the record for a period in which instrumental records are very rare, the data and results presented are relevant to the scientific community. However, in my view the manuscript presents major flaws in terms of how the data are analysed and the results interpreted.

General comments:

- The comparison with modern data is misleading, as the historical data are affected by unknown biases related to the measurement location, thermometer scale, and other factors. These biases can easily surpass climate variability in magnitude. I understand the appeal of comparing historical extreme climate events with modern climate, but this is not possible in a scientifically sound way without applying the necessary bias adjustments - which is very hard if not impossible in this case given the lack of metadata. This is particularly true for temperature and cloud cover.

- In addition to the previous point, the Meteoblue dataset used by the authors is a commercial product that, as far as I know, has never been evaluated in peer-review literature. If that is true, it should not be used for a scientific article. My suggestion is to drop the use of a modern reference period for temperature and cloud cover - except perhaps to assess data quality such as for the NDR calculation - and concentrate on the decadal variability of the studied period.

- The temperature record is clearly affected by inhomogeneities. The authors actually do a very good job in pointing them out, by mentioning relocations, changes in the temperature scale, and changes in the ventilation of the instrument (Tab. 1; also Fig. 1a and 1b point to at least two important inhomogeneities). However, this fact is completely ignored when analysing the data. There are some confusing sentences about this at the end of Sect. 3.3 that actually raise even more doubts about the quality of the data. I believe that some kind of statistical homogenization is necessary, even though reference series for this period are scarce. Beside the Central England Temperature series, there exist many temperature reconstruction that could be used.

- How does this record relate to the widely available long monthly temperature series for Paris ? Is that series also based on Morin's observations? Are there any differences from your data?

- Many equations and definitions appear in the results. They should be moved to the methods section.

- I am not a native English speaker but the quality of the language seems rather poor to me, to the point that I had difficulties understanding some sentences.

Specific comments:

- The procedure to convert the temperature readings to Celsius need to be explained more in details, since the given references are in French. Besides, the conversion formulas (Eqs. 1-3) are not completely clear to me: I would expect that the TM in the three equations refer to different observation times, but this is not indicated. Moreover, it is often mentioned in the manuscript that the the thermometer was filled with spirit: is this an assumption or a known fact? How do you explain that a linear conversion does not introduce a bias at high temperatures?

- Equation on page 10 (number missing): I believe the indices i,j,k here are in the wrong positions.

- P11, L207: Dai (2006) shows that the effect of pressure on snowmelt is negligible in the lower troposphere. Besides, increasing humidity cause the melting point temperature to decrease (i.e. a lower temperature is required for snow), not increase. More importantly, precipitation phase at the surface depends on the temperature profile above the station, of which surface temperature is merely a proxy (e.g. it can be significantly warmer 1 km above the surface than at the surface, hence it can rain with negative temperature). Another important factor is precipitation intensity (higher intensity implies higher melting point temperature).

- P11, L215: "So, if..." - something wrong with this sentence, snowfall frequency is not measured in °C.

- Equation on page 12 (number missing): What is the factor 2 for? The notation for the sums is confusing.

- P12, L231: What is a typical value for NDR for data measured indoor? How relevant is the change from 0.8 to 0.95 in 1688?

- P13, L263: The Maunder and Spörer Minima are defined by solar activity, not by climate, and the influence of solar activity on climate is still uncertain. This is mentioned briefly later in the manuscript, but I believe it should be clarified already in the introduction. The choice of LMM to describe the period covered by the data is perhaps not the best as it gives the impression that the climate anomalies were mainly driven by solar activity.

- P22, L349: "This means that..." - Circulation is an essential requirement for cold winters rather than an additional driving factor. Even a possible solar influence would mainly act through changes in circulation (e.g. Barriopedro et al., 2008).

- P22, L367: How is exactly the DI calculated from modern data? Are clear days excluded? If not, there would be an obvious bias with respect to Morin's observations. How dependent are the results from the choice of the levels? This comparison should be done using an open, peer-reviewed dataset (e.g. ECMWF ERA5 reanalysis), or dropped.

The authors have carefully examined a part of the measurements contained in Louis Morin's observation notes kept from 1665 to 1713. These notes are of great importance, given the scarcity of measurements at that time, for understanding the climate during "The Little Ice Age". Temperature and pressure measurements have been the subject of previous studies, well mentioned by the authors. We believe that the interest of the article relates to:

- the confirmation of the high quality and reliability of Morin's temperature measurements.

- the confirmation of the characterization, using these, of the climate of this period

- the study of observations of the direction of the wind deduced from the movement of clouds, which had not yet been the subject of previous studies.

This last point allows the authors to establish a link between the temperature differences (in particular the remarkably very cold winters-springs, but also the summers-autumns close to modern values) with the characteristics of the atmospheric circulations.

The abstract gives a precise idea of the content and of the conclusions of this study.

Morin's Manuscript, which can be consulted at the Institut de France and has already been used by several of the authors cited in reference, includes observations up to 1713. We are surprised that the authors of the article could not use a complete copy of Morin's manuscript or even failing that, did not use in their analysis the data from 1710 to 1713 published in volume 2 of the 1992 note by Legrand and Legoff, cited in the article.

To be more in conformity with the title of the article, it would be desirable to include these data de 1710 à 1713  in the study or otherwise to modify the title of the article in : “Subdaily meteorological measurements of temperature, direction of the movement of the clouds, and cloud by Louis Morin in Paris from 1665 to 1709”.

Some other points, indicated below, seem to us to be able to be improved:

Line 87: it would be preferable to indicate here only the Fontenelle reference. The same passage quoted by Legrand and Le Goff is taken from the original in French

Line 142 to 164. The authors seem to have completely adopted the method developed by Legrand and Le Goff, 1992 to convert Morin's measurements into °C, a method explained in 10 pages. In this article, the summary which is given in 21 lines does not make it possible to understand the principle, in particular that the 2 periods indicated in line 155 are the years of observation of Morin from 1676 to 1712 and those of the Observatory of Paris from 1816 to 1852 considered to have identical average maximum and minimum temperatures. Are the values ​​in °C recalculated by the authors identical to those published in volume 2 of the book by Legrand and Le Goff?

Line 164 and 165: “Rousseau, 2009” and not “Rousseau, 2013”

Line 166: “grape harvest dates” and not "harvest date"

Line 172: “monthly” instead of “daily”.

Line 171 and Figure 4: .

The assessment of a bias in Morin's measurements from 1776 to 1780, which may possibly be taken into account for the calculation of monthly temperatures, is different according to Figure 4 of this article or Figure 1 of the article Rousseau, 2009. It seems that   value for 1677  of harvest dates in Beaune used in the article (series of Labbé et al)  is not in agreement with the chronology of the temperatures of Morin, which is in phase with the chronology of the Dijon series (series taken from Angot) used in Rousseau, 2009.

Line 177 to 179: Writing too technical, difficult to understand. In addition it seems that this smoothing of the data does not allow better readability, the curves somewhat masking the raw data. Wouldn't a data visualization for Figure 5, 8 and 9 similar to that used in Figure 4 be better?

Line 204 and figure A1: For comparison it would be interesting to provide for the Meteoblue data from 1986 to 2015 the 2 Figures similar to those of appendix A1 concerning the Morin data.

Line 221 and figure 3: Is the same difference observed between morning, noon and evening on the Meteoblue data? Morning, noon and evening Meteoblue curves could be shown in Figure 3.

Line 235 and figure A2B: the differences noted in the thermal amplitude illustrated in the A2b would deserve comments and undoubtedly a more in-depth study (on a finer scale than the year), which could possibly make it possible to detect more precisely ruptures of the homogeneity of the series, the extreme values ​​being more sensitive to the local environment of the observation. Figure A2b seems to indicate breaks around 1680, 1690, 1700, 1705 which do not coincide with Morin's changes of domicile. Are deviations of such great amplitude observed in Meteoblue data? Are these differences related to inter-annual variability or to changes in location or others modifications of measure conditions?

igure 4 – Legend: “Beaune” and not “Dijon”

Line 276 and table 2: "the extraordinary positive anomaly" observed between 1676 and 1680, which is not found in the CET temperatures, confirms the hypothesis of a break in the homogeneity of the Morin series in 1680 which could be dicussed here

Line 285: The same analysis of cold days for the period 1665 to 1675 seems feasible despite the lesser precision of Morin's small thermometer (distribution of measurements noted f4?) even if it means homogenizing with respect to the following period, due to a different threshold.

Line 293: Same comment concerning the hot days (distribution of measurements c4?) of the period 1665-1675

Line 318: The growth of TCC does not take place over the entire period 1676 to 1709 and therefore the use of a growth rate of TCC from a linear regression established over the entire period is not actually justified. We note very clearly in figure 8 that the 5 curves would be rather decreasing or stationary after 1693

Line 393: Is the only strong deviation of westerly winds observed in the decade 1700-1710 not due to the fact that it is the only decade including a complete phase of positive temperature deviations? The other decades present both rather warm or rather cold temperature phases of multi-decadal fluctuations. It would be interesting to examine whether a division into 4 periods, corresponding to alternately cold and warm phases of multi-decadal fluctuations 1672-1675, 1676-1686, 1687-1701, 1702-1708 (cf Le Roy Ladurie et al., Fluctuations du climat, 2011) would not give clearer differences.

Line 454-455: The complete consideration of temperature data from 1665 to 1675 (winter 1672, summer 1675 remarkable in particular) and from 1710 to 1713 as well as the question of the break in the homogeneity of the measurements suggested by the figure A2b, seems to us to deepen later if the article does not deal with it.