Thank you, Reg. My reply would have been much less gracious and not
nearly so complete. Probably something like, "Which part of 'constant
DC input' do you not understand?" I appreciate that you took the time
to write such a nice reply, one that the lurkers may well learn
something from.

Another reason, of course, that a tube can't handle modulation peaks is
that it's getting weak (or the initial design was inadequate). Seeing
the plate current meter wiggle during modulation that's below 100% is a
good sign it's time to check the RF output tubes, and as your reply
suggests, the drivers as well, though the grid current meter should
tell you enough about that. If the filament (cathode) no longer has
enough emission, the tube may handle the carrier OK, but not the
positive modulation peaks where the current must be about twice as high
as with just the carrier. Then the DC current will drop during
modulation.

Cheers,
Tom