The drifting discriminator threshold level seems to have been due to a bad solder joint. The joint is between the amplifier cards "Upper V19" and "Upper V18". The voltage seemed stable at the higher "V" numbers and unstable along the lower "V" numbers, toward the target end. The joint is a piece of wire soldered on a trace. The wire bridges from one circuit card to the next. These are the circuit cards that carry the wire signals and power to the amplifier/discriminator cards. I soldered a longer jumper wire to the traces to jumper the cards' traces together. Things seem to be stable. The threshold seemed stable on the other side of the top VDC, the "U" side. The top "V" side is where we read the threshold voltage back from.
The most likely outcome from a drooping discriminator threshold is that
the amplifier might start oscillating. It is unlikely to get a decrease
in wire chamber efficiency, unless the amplifiers are oscillating. It
would take an increase in thereshold to get a decrease in efficiency.