well lots of times the fwhm will be different per-filter. if it's different enough then the plots can suffer from the y axis scale being too large and you end up looking at two lines at different heights, which can make the outliers *less* obvious.
rob