Sorry to be so anal but if you're talking about this tube:
http://www.radiomuseum.org/tubes/tube_2d21.html
I'm puzzled as to how that would work. Thyratrons are basically antique SCRs, they are triggered by a pulse, and won't turn off until the plate voltage gets down to 0. Correct me if I'm wrong, to be honest I haven't played around with Thryatrons too much.
Mmh, turn off when plate voltage falls below 10 or 20V I think, or more to the point, if current flow drops below holding current, AND enough time passes for the ions to recombine. (Grossly the same physics as in a thyristor!) If you skip that last key step, it'll keep right on conducting when the voltage comes back.
Inverters are easy to make, you just have to provide some means for commutation (turning one off and the other on). Usually, while one side is on, the opposite side is triggered, shorting out the supply, which swings around by means of a resonant tank circuit, reversing current long enough for the opposite device to turn off. I've played with
this example before: the minimum pulse width was something like a hundred microseconds for the devices I used, so you get useful PWM range up to a few KHz (i.e., a duty cycle of, say, 30-70%). There are equivalents for PP, full bridge, and supplementary commutation (i.e., an outboard device which can trigger the mechanism, resulting in all devices turning off).
Tim