my review attached. 

The manuscript deals with aerosol and cloud droplet microphysics determined at a mountain site in the Peloponnesus (Greece). This is part of a series of publications dedicated to the CALISTHO campaign. Aerosol activation in orographic clouds certainly is not a new topic, but CALISTHO convincingly provides a demonstration of what can be achieved by employing state-of-the-art instrumentation including wind lidar observations for assessing vertical velocity regimes. The paper provides a comprehensive set of relevant new information about emerging ground-based observations targeting aerosol-cloud interactions. The manuscript is generally well written; however, Section 3.4 deserves more focus as it contains some obscure arguments.

Specific comments:

Section 2.3.1: When determining the hygroscopicity kappa value starting from the chemical composition, what factor has been applied to the organic mass fraction?

Lines 301 – 303: “we identified three prevailing wind directions, that correspond to the local transport patterns (Figure S2f) from 90°, 180° and 320° N, where the NTotal obtains its maximum values (~3300 cm^-3)”. It is not clear whether the maximum values for NTotal are found only at the 320° direction or all the three selected wind directions.

Line 306: It is not really clear in Figure S2 what are the wind directions tracing an air flow passing over mountain peaks. In addition, with wind directions from 330 – 360°, the PBL height is large and vertical velocity as well (Fig S2b) which is the other way around with respect to what stated in the text.

Lines 308 – 310: the references to the panels in Fig. S2 (panels b, c, d) in the text do not match with what shown in Fig. S2 in the supplementary. Are the Authors referring to Fig S3 instead?

Fig. 2. Panel b: the PBL / FTL mask is not clear; for instance, in days 11-14 Nov the PBLH is lower than the (HAC)2 elevation but this period is classified green (PBL); the same occurs on 20 – 23 Nov. Panel c: In the wind direction plot, please use a palette with colours changing with continuity between 360° and 0°.

Fig. 3a: the y axis of the figure on the left cannot report simply concentrations in cm-3 but must be in dN/dlogDp units. The specific period of the campaign providing this subset of data should be reported. Same for Fig. 5.

About the “virtual cutoff”. Clearly, a (physical) PM10 cutoff cannot provide a net separation between interstitial aerosols and cloud droplets when Deff is small, which in fact results into too small values of the difference [pre-cloud]-[in-cloud] concentrations with respect to CDNC (Figure 3b). The interesting analysis performed by the Authors leads to the conclusion that residual (evaporated) droplets are absent in the PM10 inlet when Deff is larger than 13.5 µm (line 363 – 364: “Thus, we confirm that this threshold is consistent with allowing only interstitial to pass through the inlet”). This would imply that measured interstitial aerosol concentrations are reliable only for cased with Deff > 13.5 µm (which represent a limited subset of the full in-cloud measurement period, Fig. S4). However, the following analysis (Fig. 5) is carried out on an “interstitial aerosols dataset derived after removing from the in-cloud dataset the data were the Deff was exceeding the threshold of 13.5 μm” (lines 431 - 433). However, with Deff < 13.5 µm, the interstitial dataset should be contaminated from the concentrations of droplet residuals. Please, explain.

Section 3.4: if clouds are observed only at the (HAC)2 site and not at VL, why the values in the y axes differ between the two graphs in Fig. 6? Most noticeably, in the FTL conditions, Deff is larger than 13.5 µm and the PM10 inlet at (HAC)2 site should sample interstitial aerosols free of cloud residuals, so why the closure is worse in this case? Please, clarify.