Hoi, the following contains scans for mm2 and mx with the parametrization of the theoretical error as currently implemented in the fit (i.e. determined with generated MX). The study cutting on the fitted Mx will follow: http://www.slac.stanford.edu/~ursl/talks/062002/mx-stability.eps http://www.slac.stanford.edu/~ursl/talks/062002/mm-stability.eps The mx scan does not reveal anything new compared to Daniele's plot, except that the theoretical error is added in quadrature (this is the outer error bars, the inner error bars are just the statistical error). The numbers for the points are contained in http://www.slac.stanford.edu/~ursl/talks/062002/mx-stability.txt http://www.slac.stanford.edu/~ursl/talks/062002/mm-stability.txt Here, the errors in parentheses are the combined absolute and relative errors (on BRBR), respectively. Cheers, --U.