I did not run the amp continuous with a load on it. I did do load testing for THD, frequency response etc. The fact that I could consistently see the bias current run away on one tube, and replacing that tube stabilized the bias current, seemed conclusive that the tube was at fault.