Hi, I repeated the overlay of the fit results in the subsamples that showed the largest discrepancy between Q0 and Q1. superblocks: http://www.slac.stanford.edu/~daniele/vub/sbcomp.eps (top left sb1, top right sb2, bottom left sb3) run 1 2: http://www.slac.stanford.edu/~daniele/vub/runcomp.eps (left run1, right run2) in each subsample you find the usual third bin off. In run 1 the effect seems to be more evident. Daniele