I discovered something interesting today. Yesterday, I was measuring plate voltages on the 5881's in my Bandmaster Reverb, as part of dialing in the Bias. On one tube the plate voltage was 466. Today, when I resumed testing, without me changing anything, the plate voltage on the same tube was 448. I thought about it for a while, and considered what is different today? It's about 30 degrees cooler this morning and I have a little ceramic space heater on in my office, on the same circuit the amp is plugged into. Sure enough when I turned off the heater, the plate voltage went to 466, on half power on the heater, 458 volts, and at full power, 448 volts. Yesterday I was measuring idle plate dissipation on that tube of 80%, Today it is 67% (The voltage drop and plate current are both considerably lower). I guess I had never considered that other devices on the same circuit can significantly affect the amp biasing process.
The wall voltage on that circuit varied from 122.3 with the ceramic heater off to 117.7 with the heater on full. I also have my computer and a halogen desk lamp on that circuit.
I though this might be interesting information for anyone who messes around with tube amps.