X-axis during multipeak fit

Hi guys,
I am trying to do multipeak fit for the amplitude histograms of EPSCs. I start by using "Autoloscate peaks now" button. In principle it works, but it seems that location of the detected peaks and sometimes even their number depends on the x-scaling. I do not change anything with the amplitude histogram itself (= I do not change number or width of the bins), only set the x-axis ranges differently, and this changes the result of fitting. Somehow, it does not make sense to me. Can somebody explain why does it happen and what is the correct way to set x-scaling then?
Thank you,
Maria
Could you post an Igor experiment file with the data, along with details that will let me try out exactly what you have done? Also please include details about what you expect, and what happens that is unexpected. Thanks!

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com