Let me know your thoughts on this:
In many old, transformerless, series filament (AA5) radio/amps, one side of the AC is indeed tied to the chassis, along with the ground for B+ ...
In such a case, is it possible then that the B+ on the heaters would be carried through the heaters?
This assumes that the cathode really is tied to one side of the heaters, which I havent fully verified, but seems likely.