High voltage probe, odd NST measurements

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Michael S <mike5904@xxxxxxxxx>

I just built a simple resistor voltage divider for the purpose of
direct measurement of high voltages in the 1-50kV range.  It consists
of 10 1Gohm resistors in series with a 1 Mohm resistor.  Voltage is
measured across the 1 Mohm resistor, which should create a 10000:1
divider.  I decide a simple test would be measuring one of my NSTs, a
12kV/60mA Allanson unit.  Using a fairly disposable meter I had around
I hook up the low voltage end of the probe, connect the probe across
ground and one of the transformer legs.  I apply about 50V to the
transformer, and get no reading.  Confused, I try to draw an arc off
of one of the terminals to a wire connected to the transformer case
(not the best idea), which works fine.  Now even more confused, I
decide to switch the terminals on the probe.  I insulate the meter on
a sheet of PVC first, since it will now be at the potential of one of
the NST legs, and power the transformer up again.  Now I get a reading
of .55, which would correspond to a voltage of 5.5kV from leg to
ground.

Here's what I'd like to know:
-Why do I get no reading when measuring from ground to leg, when there
is obviously a potential difference there?
-Why do I get a completely different reading when reversing the probe
(and putting the meter in an unsafe position)?
-Why is the second reading twice that of what I would expect for
ground to leg? (since I am only applying about 50V)

Thanks for any help.