Follow up to tell folks the current status and probable cause of the problem
(I think):

I replaced the old 10nF ceramic smoothing caps in the grid bias supply as
they were getting distinctly warm after the scope had been on suggesting
that they were getting rather lossy (on a damp day they would steam gently
for a minute or two after turn on).

Didn't solve the brightness problem but new polyesters do NOT get warm, so I
feel better :-).

Static grid voltage at output of grid cut-off adjustment trim-pot was
drifting up during warm up, while voltage on input side was pretty much
constant. Suggests it might be the trim pot drifting with temp.

The trimpot (R315) is supposed to be a 2M2 linear Cermet 20% 1W according to
the parts list, but looks to me like a standard Piher PTC15 1/2W part.
Measuring the resistance as the scope cools after power off seems to confirm
the hypothesis (10% increase in value over 15 minutes, probably as much as
15-20% by the time it reaches room temp).

If I could source a replacement readily I'd just ram a new one in, but I
suspect that the parts list is right and the original WAS a 1W part, and has
been replaced somewhere along the way with a 1/2W part. Finding a 2.2M
0.5W trimmer of any sort (never mind one that will fit) seems a bit hard,
and as for a 1W part ...!

Before I go totally insane hunting for a suitable part, I'd just like to get
a sanity check on my diagnosis from folks more experienced with this than I.

Thanks
Dave