This paper investigates the growth of ice particles through the Bergeron-Findeisen process by using Direct Numerical Simulations (DNS).  The paper is well written, and the results are interesting. I can recommend this paper for publication, but I would like the authors to answer next questions:

1) I would like to know more details about the simulations, such as grid spacing, time stepping and used numerics. It would be also interesting to know if the droplets/ice sediment, and with which velocity. 

2) In the description of the measured environmental conditions (table 1)  it would be useful to have the mixing ratios qc und qi and the number concentrations qnc and qni. It is also said that drizzle was measured, but I guess this has no influence on the experiments.

3)  The DNS experiments have a dissipation rate of 10m²s-³. Is this not too high for such a cloud?

4) Please state the mean temperature in the experiments.

5) Why do you use a standard ice concentration in the DNS of 1000 l^-1 when only 1l^-1 was measured?

6) The experiment RH_noGC starts with unusually low radius of droplets (R=1.5 μm), which I do not think can be easily found in clouds. It is then not surprising that all liquid water quickly evaporates if RH<100%. Please motivate these experiments. 

7) It is a common assumption in weather models/LES with a time step 1-20 seconds that liquid water is in thermodynamic equilibrium, but not the ice. The underlying assumption is that water condensation/evaporation is much faster than ice processes. Could the authors comment on how good this approximation is? These experiments could be very useful to verify the above-described approximation.

Please find my review attached below.