[replying to two messages in the same thread together]
   I was
using a 40W light bulb in series with the primary, as a
 current limiter, which is why the primary voltage was 25V instead of
 ~120V.  The bulb was glowing (not quite full power, but close). 
 Well, even with a
40W bulb in series, I would expect a transformer
 with no load on the secondaries to give almost full output if there
 were not shorted turns. 
 
 Well, yes - since it would be drawing (to a first approximation) zero
 current.  That the bulb glows at all with all the secondaries open is 
Not so. There will be an out-of-phase current due to the inductance of
the primary winding. And that current, if sufficiently large, would cause
the lamp to glpw
_But_ my experience suggests that for small-ish power transformers, the
sort of thing we're talking about here on no-load, a normal mains lamp in
series with the primary will not glow. You could trie a higher-wattage
bulb, thoguh.
IIRC, though, you've got a leak (or short) to earth from some point in
the HV widning, and another point on that winding (the centre-tap of the
B+ part) is deliberately earthed. That can't be right. And I'll bet that
short-to-reath is not just on one turn, in other words you also have
shorted turns there.
    Turns out the winding I thought was the HV winding,
only shorted
 out, is actually the CRT filament/heater winding.  (I'm quite sure
 I've got the CRT filament pins correct; they are the only two CRT
 pins that show DC continuity.) 
 It's always possible (you'd better hope
not!) that there's an
 internal inter-electrode short in the CRT. 
 
 Unlikely.  Those two pins show about three ohms resistance.  I'd expect 
 
SOunds like the heater...
 Then it probably is a heater, with that odd circuit serving some other
 purpose.  (There is a "Z input" on the front panel, which is a
 brightness control; it is capacitively coupled to this possibly-cathode
 pin, which would make some sense....) 
Sure. The Z-modulation input is an external brightness control input,
used for some special applications. In general it's capacitively-coupled
to either the cathode or control grid of the CRT.
   The HV
winding appears intact, and runs between one side of the B+
 winding and the filament of one of the HV rectifiers; 
 That suggests to me a -ve
EHT supply (output taken from the anode of
 the rectifier). 
 
 I now believe there is a -ve EHT supply and also a +ve EHT supply.
 There are two HV rectifiers; if I draw the rectifier valves as
 semiconductor diodes instead (to simplify the ascii-graphics), omitting
 their filaments for simplicity, the circuit looks something like:
  ===================
 .oooo.oooo.oooo.oooo.                   +--> resistor chain (see below)
 |    |    |    |                        +--> one end of CRT heater
 |   GND   |    |               100K     |   220K
 |         |    |           +--/\/\/\/---+--/\/\/\/--> CRT cathode(?)
 |         |    |    +--|<--+            | .5uF/2KV     47
 +->|-+-|<-+    |    |      +---||--GND  +---||---+--/\/\/\/--GND
      |         +----+      .5uF/2KV              +--/\/\/\/--3.15VAC
      |              |                  .5uF/2KV        91
      |   ====       +-->|---/\/\/\/--+-----||----GND
      |  .oooo.               100K    +--/\/\/\/--GND
      +--+    +--+                    |     5M
                 |                    +---------> CRT final anode
                 +--> B++   +--> B+
                 |          |
       GND---)|--+--/\/\/\/-+-| . |---GND
         80uF/475V             0B2
 (The unconnected final winding of the transformer is actually the
 filament winding for the rectifier whose cathode is shown connected to 
 
^^^^^^^^
Don't you mean anode (plate to you) here?
  one end of that winding.) 
  Add in the four deflection electrodes and this
accounts for all but one
 of the connections to the CRT.  That one is driven by a circuit I do
 not understand which appears to have something to do with the "Y LIN"
 internal adjustment - this may make more sense once I've traced more of
 the circuit. 
It may well be an internal shield between the defleciton plates. Altering
the voltage on it wil lcontrol the shape of the beam-deflction .vs.
voltage between the defleciton plates characteriostic, thuse it will
LINearise the Y defleciton.
-tony