I'm afraid there is no way to reliable reverse calibrate. The only way to get an approximation would be to prepare the 1986 curve data in the appropriate form for OxCal (it is not supplied with this) and then try dates to try to get the appropriate range. However, I'm suspicious of the ±85, since calibrated ranges are normally asymmetric. At this date they will have been done as an intercept plot - and you don't even know if this is one or two standard deviations (though probably 1 from the value).
Christopher
> --
> You received this message because you are subscribed to the Google Groups "OxCal" group.
> To post to this group, send an email to ox...@googlegroups.com.
> To unsubscribe from this group, send email to oxcal+un...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/oxcal?hl=en-GB.
>