Keithley 6514 gives odd SCPI read response

[adamredwine]

I am using a Keithley 6514 to measure some small currents (~100 pA - 0.5 microA) in our lab.  Because these values cross several orders of magnitude, the electrometer shifts among the different ranges.  When it does so, I sometimes (though not always) get a step change in the readout provided by a SCPI query.  That is, over a series of several measurements, one section will appear reasonable and at some specific point (when the range changes), the measurements will jump by a factor of 2 and then the subsequent measurements will be consistent with this new baseline.  Has anyone else seen behavior like this?