I ran the MIP calibration over several Emc Check runs from 2008.  I used the same MIP dE/dx and sampling fraction as for past calibration, derived a calibration based on MIPs and compared to what is in the database. I am finding gains that are consistent with the database values.  Attached is a ratio plot that compares the values I fit to the database values.  I added a fit to a straight line.  Unfortunately for the first one in the plot the fit is getting pulled low by one stray value, but the rest are mostly consistent with 1.  Below is a plot of the fitted gains compared to ideal and you can see there is no systematic shift low.  The scatter seems to be due to the general instability of the fit used.  This is one reason I think we should just use the histogram mean and not the fit.

For the record, the runs used were Emc Check runs between 9001018 and 9070003 from run 8.

I also attached a comparison of run 9 fit values I am getting to the run 8 values in the database.