I found what I was looking for in Roger Lapthorn's notes on
earth mode at

  https://sites.google.com/site/sub9khz/earthmode

where he mentions that licencing is not required.  However he
is talking about earth mode where comms are via ground currents
and there is no intended use of far field radiation.

This is in contrast to an earth antenna where the ground
current forms the lower leg of a vertical loop, such as is
used by the military at ELF.

Last night I fired up the amplifier and through the transformer
was able to put about 600V RMS onto the wire, so about a
kW into 380 ohms.  The transformer worked very well but
for some reason the amplifier kept wanting to do a thermal
shutdown which is odd considering its 5kW average power rating.
It's supposed to do 2.5kW average into a 1 ohm load.

The test was at 8270.01 Hz but not even the faintest trace
showed up on Gary G4WGT's spectrogram at 35km distance.  Not
surprising really, the ground around here is quite unsuitable
as an earth loop: a metre or two of wet soil over the sandstone
bedrock, the return current is not going to go deep.

The transformer appears quite good at 8270 Hz and looks like
it might work well at coupling power into the base of a VLF
loading coil.

The ground antenna has so far qualified itself as a nice high
power dummy load.  This evening I'll play a little more
with it just to understand what the amplifier is doing but I
don't think I'll continue much more with these earth experiments.

David Hine wrote:
> take care as the mains 'earth'

Yes.  Several years ago I experienced just that.  I found, I think,
about 40V between the domestic earth and a VLF antenna earth.
Eventually I traced it to an electrical fault in a neighbour's
tractor shed.  The entire domestic ground of our little
cluster of farms was at significant mains potential relative to
a more distant earth ground.  This might have gone unnoticed had I
not been setting up a remote rx with its own ground.  During a
storm, nearby lightning could easily produce very large transient
voltages between domestic ground and the ground of a remote antenna.
I always disconnect everything on the approach of a storm.

Right now, about 1.5V RMS at 50Hz between the ground 500m away and
the house ground.

It isn't unknown for farmers to run a single wire carrying live out
to a distant shed with the return current via the ground.  If anyone
was doing that near here I would soon know about it!