[phenixbb] Which is the better way to combine the dataset from different crystals?

[fn1 at rice.edu]

Hi all,

I have a question on how to combine two dataset from different crystal  
of the same protein.

The first way I could think of is that,
1) Merge two dataset seperately;
2) Scale them to see whether they are consistent with each other  
enough as indicated by the R-factor;
3) If the R is low enough, say below 10 or 15%, then take the average  
for both the F and SIGF.

The other way is that,
1) Take the unmerged file from MOSFLM, and reset their batch number;
2) Run SCALA to scale these two unmerged dataset at the same time.

Could anyone tell which way is the better or the correct way?
Thank you in advance!

Alphar