In general the Rvalues of course fit very well Rmerge. A good sign that it should be fine, as far as I remember correctly about the theoretical limits for the Rvalues after refinement compared to the experiment. Best Regards Christian