The founding claim: a voltage field pulls the water molecule apart. In the bulk it cannot — it turns it. But almost none of his voltage is in the bulk. It is in the last three tenths of a nanometre against the steel, where the field is within a factor of sixteen of where water comes apart.
Every circuit, cavity and injector in this archive exists to make one sentence happen. From Electrical Polarization process:
Stationary "positive" electrical voltage-field … not only attracts negative charged oxygen atom but also pulls away negative charged covalent electrons from water molecule. At the same time stationary "negative" electrical voltage field attracts positive charged hydrogen atoms.
He calls it the Electrical Polarization Process, and figure 3-26 on the same page draws it, down to the words covalent breakup.
In the bulk, a field turns a molecule. It does not stretch it.
A water molecule is neutral — a dipole, charge apart rather than charge missing. The field pulls the two ends with equal and opposite force, so the net pull is the difference between them: it depends on how fast the field changes across the molecule, not on how strong it is. In a perfectly even field, all you get is a turn.
But his field is not even, and he knew it. Figure 3-29 draws the field lines bending around one electron, labelled with an attraction force and a repelling force. And Resonant Action gives the cell: "typically 0.50 inch diameter tube inserted into 0.75 inch diameter tube having a .0625 concentric air-gap 3 inches long" — coaxial, so E goes as 1/r. There is a gradient. The effect it produces is real and is used daily (dielectrophoresis, Pohl 1951).
What is wrong is the size: at the inner tube, at 1 kV, the net pull on one water molecule is about 7 × 10⁻²² N against a thermal scale of 4 × 10⁻¹² N. Ten billion times too small.
The bond is electrostatic, and he draws it correctly
The usual correction — "a covalent bond is not an electrostatic force, there is nothing to pull on" — is wrong. By the Hellmann–Feynman theorem the force on each nucleus in a molecule is exactly the classical electrostatic force from the electron density and the other nuclei. The bond is the nuclei attracted to the shared density between them; quantum mechanics decides where that density sits. Meyer's "electrical attraction force (q, q′)" on figure 3-27 is nearer the textbook than the correction.
And his own later chapter does not ask for a tear. Covalent Switch-Off:
applied Opposite Electrical Stress … directly weakens the covalent bonding of the water molecule (q-q′)
Weakening is a different, and testable, claim: a field shifts the O–H stretch frequency. That is the vibrational Stark effect, and it is measured with infrared.
Where the voltage actually goes
the gap in this cell
0.0625 in — 1.59 mm
which is about
16 million O–H bonds wide
so at 1 kV, one bond's share
60 microvolts
against the 5.15 eV it wants
short by 85,000×
That is the honest number for the bulk, and it is worse than the figure usually quoted. But almost none of his voltage is in the bulk.
At the plate, every number changes
Amp Leakage found that almost all the voltage across a water cell lands in a layer about 0.3 nm thick against each plate.
Within a factor of sixteen. His voltage was arriving exactly there the whole time. Which also means field-driven dissociation and electrolysis are not two doors: electrolysis proceeds through that layer. Nicholson and Carlisle, 2 May 1800 — not 1799.
What "tune in to the dielectric property of water" means
Not the molecule. Resonant Action, the next page of the same memo, defines it:
allows voltage intensifier circuit … to tune-in and match the resonant characteristics or resonant frequency of water bath … The established resonant frequency is most generally in the audio range from 1 kHz up to and beyond 10 kHz; and is dependent upon the amount of contaminants in natural water.
It is an L-C resonance: the chokes against the cell's capacitance, and the water sets the capacitance. Which is checkable, because both halves were metered. Two estate chokes (≈2.4 H) with Gabel's cell:
water in the cell
C at 1 kHz
resonance
distilled
25 nF
655 Hz
rain
21 nF
715 Hz
tap
5.72 µF
43 Hz
Clean water lands just under his stated band; tap water drops it off the bottom by a factor of fifteen — dependent upon the amount of contaminants, exactly as he says.
(A water molecule does have a frequency of its own — 19 GHz, which is why a microwave oven runs at 2.45 — and it is four million times above where he drives. He never claimed it.)
The score
water polarizes in a field
right — that is what makes it 78.5
voltage is the thing to apply
right
his pulse rate is fast enough
right, by 12 million times
his cell's field has a gradient
right — coaxial, and he drew it
a field can weaken the bond
right, and infrared can see it
it happens in the bulk of the water
no — it happens in the last 0.3 nm
One no, and it is the one that rescues the other five.
What would settle it
Coulombs in with a current probe into a scope, not a panel meter — an average-responding meter under-reads a narrow unipolar pulse train badly, and will make an ordinary electrolyser look "over Faraday". (The 1993 Southampton account in this archive, gas at under 1 mA melting stainless wire, is very likely that error: 1 mA for an hour is 3.6 C, 0.42 mL of hydrogen and 5 J.)
Then gas out over water. Then — separately — joules in against joules out, because that is the claim he actually made. Faraday's law can be obeyed and Meyer still be right about economy; the coulomb count alone cannot tell them apart.
See Gas yield and Faraday efficiency, Coefficient of performance and Water permittivity at temperature.
