I think I have an answer here.To end the argument if this is a battery or not.
Put a layer of water soaked paper between two glass slides .Seal the edges with a waterproof glue.Let dry.Then perform this experiment.If the water is confined to inside the two glass slides then its not wetting the two dissimilar metals.Then the diamagnetic properties of the water layer should be apparent.I'm not able to do this experiment.But maybe someone else  here is?And report back here?Triffid

Oops,looks like i goofed.Some of the guys here have gotten millivolt readings without the water being present.When the water is added.I am more interested now in this magnet battery now than I was before .A simple LRC circuit could be designed to increase the voltage at least 700 times.triffid

I meant to say when the water is added.you get the higher volts because of a battery action.I have a good friend who has several u shaped magnets who intends to do
some experiments along these lines.It will be interesting to see what he finds.Also I have already designed a LRC  that turns .001(1mv) volt into 729.87mv.Of course the amps decrease in similar fashion.triffid

Okay, here's for a bump and some more thoughts...

So let us assume for a minute that it really is possible to
pull a current from the 'potential difference' created when a
conductive permanent magnet has a flux path through a
"keeper" that is electrically isolated from the magnet, while
this isolator has a high magnetic permeability so the flux does
get through but the electrons just can't.
At first glance I would say that such a setup using a horseshoe shaped
nickel coated neodymium magnet with a keeper and a paper "dielectric"
in between them should be replicable in a setup where 'normal' block
magnets are used, as long as the two block magnets are equally strong.
Also it does not seem farfetched to consider replacing the nickel coating
with a block of nickel (or any other magnetic metal) placed in between
the magnets... after all, if the current thus produced in the wire is indeed
a result of the Leedskalninesque quasi-monopole particles flowing from the
center of the magnet part to the center of the keeper simply because those
same pole flows can't pass the dielectric for some unknown reason and the
"pressure" of magnetic particles in the center of the magnet is so much
higher than that in the center of the keeper that the magnetic particles are
forced to flow through that wire, then it shouldn't really matter if the rest of the
magnet is coated with a conductor... as long as those central parts are conductive,
and of course also magnetic. (Not necessarily magnetised, but magnetic.)

So, in other words, the functional elements and zones of such hypothetical setup
should remain when we replace the horseshoe magnet with two equal strength
block or bar magnets connected via a magnetic conductor on one end, and connected
via a layer of electrical isolator (aka dielectric) to a second magnetic conductor,
both conductors connected at their centers via a conductive (non-magnetic) wire.
Or at least, it seems to me that it should work if the horseshoe version really does work...

Why would we want to build a block/bar version and not just use a horseshoe?
Well I'd say that's obvious: horseshoes are a horrible shape for stacking.
If we want to increase the output, the easiest method is the tried and tested
"battery" approach, by which I mean a unit consisting of several single cells
connected together.
If we could make such a battery using block or bar magnets, we could stack several
layers of these atop eachother, all of them reinforcing eachothers magnetic field
and forming one neat magnetic pile, and have their output accumulate.
If that works, then it doesn't even matter if you can only get 0,5V@2,5mA from one
of those 'cells', as long as you can make them thinner and stack them.
I imagine a very thin wafer of permanent magnetic material as the basic magnet
layers, and thin metal plates and thin plastic in between...

There are still some points I'm unclear about though...
1) What happens when you replace the thin pieces of dielectric (the paper and plastic bag)
with some thicker dielectric that does allow for most of the magnetic flux to permeate?
So for example pieces of wood or styrofoam of like 5mm thick?
Does the "magnetic battery", either the horseshoe version or the balls-in-a-row version,
still produce any output when this is done?

2) If the thing works according to Leedskalninesque monopole particle flow "theory",
then there seems to be something awry here; Eds monopole interpretation of current
clearly is one of two seperate N and S monopole flows that rotate (or "spiral" if you will)
in the same direction, but they flow (or move if you will) in opposite directions.
Since this "magnetic battery" idea uses both magnet poles equally, and the current flow
apparently only occurs between the central parts of the magnet and the keeper
('central' as measured from the ends of the body of matter where the magnetic poles are),
it would seem that any "pressure difference" between the magnetic particle density inside
the magnet and that inside the keeper would be equal for both the N and the S particle
streams, and that should only result in a flow of both N and S particles along the wire
from the central part of the magnet to the central part of the keeper, where these flows
could then split again and go their own ways. Ergo, in this interpretation, the N and S
particles should flow in the same direction through the wire. And obviously that
is in direct contradiction with Ed Leedkalinins own statements about the monopole flows
in a wire that "sees" an electric current.
In summary, it seems the observation of current flow in the horseshoe "magnetic battery"
does not accord with Leedskalnins "theory" at all! 

3) assuming it would work, could we use ferrite permanent magnets in such a setup?
Obviously they are not electriclally conductive, but they are magnetic, so we might
be able to use them as both dielectric and magnet? I imagine something like a neodymium
magnet seperated from its keeper by ferrite magnets? I'm not sure, it probably won't add
much if anything at all, but the difference between the internally conductive neodymium
magnets and the internally nonconductive ferrite magnets seems important in this,
in my opinion still quite hypothetical, effect. Could it be that the internal conduction which
gives rise to the non-permanence of neodymium magnets is also crucial in getting
any such "magnetic current" to flow? In other words, that any current pulled from such
a "magnetic battery" is in fact real-time demagnetisation of the neodymium magnet and
corruption of its internal structure?
That might explain why the tiny currents are measured in the horseshoe- (and other) neodymium
magnet setup(s), but not in real ceramic permanent magnet setups... ??

4) Has anyone determined what role the different materials play yet? I mean: has anyone tried
to use a nickel keeper with their nickel-coated neodymium magnet, instead of an iron keeper?
Has anyone tried a silver- or gold-coated neodymium magnet with the iron keeper and with
other materials? Has anyone tried wrapping the neodymium magnet in aluminium foil, or in
copper foil, to see if it still works that way? Has anyone tried wrapping a ferrite, alnico, or other type
of magnet in nickel, zinc, iron, or any other metal foil and then tried to replicate the "magnetic battery" effect?

Well I just received a bunch of fesh new neoymium magnets
of which I can be fairly sure they're of equal strength,
with which I repeated a few of the basic steps of this
"magnetic battery" idea,
and I get clear indications that there's something galvanic going on...

I tried different variations of the basic setup, used nickel coated
neos, gold coated neos, and both, I tried 2 different types of paper
(normal white printing paper and brown paper bag stuff) and 3
types of plastic which I tested for static charge retention/absorption
and found them suitable due to very low to near zero static.
I tested the neo's nickel or gold coating against aluminium in the
magnetic field (stuck between two magnets), tested it with one layer
of isolator/dielecric, with two, and with a double layer even.
I also tried one layer of aluminium without isolation vs one layer
with isolation, tested the aluminium layers respective to eachother,
as well as the nickel and gold versus the aluminium at various points
on the magnets.
I even tried a stack of isolated aluminium layers as a sort of capacitor
sandwiched in between the magnets, with and without direct contact
between the nickel and the aluminium on the North, South, and both
'side' of the magnet stack, I tried various setups with the nickel
and gold coated magnets mixed, and even only the gold and nickel
magnets with various types of isolator in between.

I did get extremely low DC output around the range between 0.5 to
50 mV briefly, but the output always dropped quite fast to reach zero,
and I was only able to get those readings using paper as a "dielectric".
The thicker the layer of paper, the lower the readings.
When the paper was replaced with plastic the meter read zero.
One nickel coated neo stuck to one gold coated neo with paper in between
gave something like 3 mVthat quickly dropped to zero, as I increased the
thickness of the paper layer the voltage decreased even more, and with plastic
it was naught.
Seems very galvanic to me...

Any suggestions?