Beste Niels,

Here some comments.
If you do not agree or consider it overdone, please remove that part.
Best regards,
  Tjeerd


Abstract, line 1:
See comment on line 223.
          line 5:
A part of the text seems to be missing between "and" and "The effect".
          line 8:
Replace "@ 68\% CL" by "at 68\% confidence level (CL)".
          line 12:
Remove the ugly statement ", we reexpress", and write 
"The fit value at the 95% CL interval is gamma = 72.9 (+17.3/-20.5) degrees".
The choice for 72.9 degrees is not obvious at 68.3% CL.
Change it also in the Conclusions. 


Line 19:
Replace "robust" by "BtoDK" or "non-suppressed".
It is not explained why "BtoDK" should give robust results
and "BtoDpi" not.
If the name is given due the shape of the CL distribution of
the full fit, then this is a biased qualification. 

Line 183:
We do not agree with the lower 68% CL interval limit of gamma in Table 5, as
extracted from Figure 6. It should be around 70 in stead of 74.1 degrees.
 
Line 223:
"The currently available tree-level LHCb measurements" is slang and does
not give precise information about the observables.
The "auxiliary inputs" is also unclear, I suggest "supplementary information".
Therefore, refer to Table 2 and 4 and replace
"In conclusion, all currently ... to constrain the angle gamma" by
"Observables measured by LHCb that are sensitive to the CMK angle gamma
at tree-level and supplementary information from other experiments 
which are given in Table 2 and 4, are combined to determine an improved 
constraint on gamma."

Line 225:
Replace "The effect of D0-D0bar mixing which can shift gamma by ~~ 1^o, 
is taken into account, considering the ... individual measurements" by 
"The effect of D0-D0bar mixing on the decay rate is taken into account
with consideration of the experimental ... individual measurements".
The information on the shift of gamma by about one degree is not important
enough to be repeated in the conclusions. It is also formulated incorrectly.