Too many data points will cause drop of curves in curve selection for plots
Robert schofield found that with 48 channels x 3 or more averages, the drop down list for selecting curves for a plot would not show all curves in the measurement. It seems to drop then in a first-in, first-out order, so that those curves earlier in time are the most likely to be missing.
It's obvious in 48 channels x 5 averages, where no curves before the third average point appear.
His file used to recreate this is something like "/ligo/home/robert.schofield/lhoDec/02-HighMifrosesih.xml" This name is close, but probably incorrect.