[sparkydave]

Thank you for the links, especially the Nostalgia Air one. It does appear that the screen voltages on the first 3 tubes are significantly high, though not because I'm measuring with a DMM. Yes, old analog meters had lower input resistance that could cause lower readings when there was significant series resistance in the way, but not 36 volts lower. I checked the 51k resistor (#47) in the circuit and it measures nearly 87k ohms. Well, that would certainly drive the screen voltages up. If the oscillator is so temperamental that a little bit of moisture will make it stop, I don't doubt that a high screen grid voltage is diverting current away from the plate and changing the operation of the tube.

Thanks again, hoping I found the problem!