Thomas Townsend BROWN
Office of Naval Research File 24-185 (15 September 1952)
An Investigation Relative to T. T. Brown
Willoughby M. Cady
1 ) Introduction
This report summarizes a technical investigation of certain
claims of Thomas Townsend Brown concerning a proposed method of
propulsion for aerial vehicles. The investigation, which began
May 3, 1952 and which is concluded hereby, is mainly analytical,
no allotment having been made for an experimental program.
2 ) Brief Description of the Alleged Phenomena
Mr. Brown claims to have discovered several new physical
effects, and considers them to be interrelated. We will be
concerned here with those relating to the force upon and between
electrically charged bodies in air or under oil.
2.1 ) It is asserted that the force exerted upon a
negatively charged body by a positively charged body is greater
than the force exerted upon the latter by the former; that the
force differential is proportional to the mass of the dielectric
inserted between the bodies; and that the phenomenon is evidence
of a new "electrogravitic" effect. Paragraph 4 contains a fuller
description of these phenomena.
2.2 ) A corollary effect is the force tending to move a
metallic disc which by means of a plastic disc supports a fine
wire, AA tangentially disposed in its plane, as shown in Figure 1. When the wire is positively
charged (e.g., 30 kilovolts) and the disc negatively charged, a
propulsion force of a fraction of an ounce appears which moves
the disc (and with it the wire) in the direction of the arrow C.
These effects have been witnessed by the writer. Further
discussion is presented below in paragraph 6.
Figure 1: Demonstration Flying Saucer
2.3 ) An apparatus has been examined superficially, which
shows an unexpected torque on an oil-immersed disc when charged.
This, according to Mr. Brown, is arranged to record
automatically and frequently the magnitude of the angular
deflection of the disc. These records have been tabulated and
partially analyzed by Mr. Brown. Greater detail regarding this
device is given in paragraph 5 below.
2.4 ) There has been briefly inspected a third electrical
device, purporting to be a communication equipment operating on
waves related to the electrogravitic effect. It appears that the
transmitter is simply a relaxation oscillator consisting of a
high tension DC power supply, a high resistor, and a condenser
which periodically discharges through a short gap. The receiver,
shown in Figure 2, consists of two
Pyranol capacitors 0-5 mf, two titanium oxide capacitors 500 mf,
a Brush recorder, and a bypass capacitor 0.1 mf to prevent the
recorder from oscillating. Two terminals in the circuit are
charged as indicated before reception and left floating during
reception. The bridge of four capacitors is housed in a cabinet,
indicated by a dashed line, which forms an obviously imperfect
shield against electromagnetic radiation. When the transmitter
is operated in an adjoining room, the Brush recorder jumps with
each spark. Mr. Brown states that the effect is present, but
reduced, if the 25 kv potentials are not applied. If the
Pyranols are short-circuited and removed, the effect is still
reported present. No attempt has been made at conscious
electromagnetic shielding of either the transmitter or the
receiver. In the writer's opinion the response may be the result
of a chance rectifying contact.
The relation between the communication device and the
mechanical effects which form the topic of this report appears
to remote that no further mention will be made of the former.
Figure 2: Electro-Gravitic Communication
3 ) The Investigative Method
This study is limited to the technical claims of Mr. Brown and
is not concerned with matters of personality or personnel, or
with the Townsend Brown Foundation. Although propulsion is the
immediate center of interest, the claimed basic effect and its
several manifestations are of the essence also. Therefore the
analysis is not restricted to the propulsion apparatus.
Rough quantitative data have been taken by the writer on the
force between charged bodies, and on the force on the disc shown
in Figure 1. He has also carried out
a brief examination of the recorded data on the oil-immersed
disc. Whenever possible, word-of-mouth data have been ignored.
It is regrettable that during the time of the investigation the
greater part of the experimental data gathered over the years by
Mr. Brown are unavailable, being in storage elsewhere.
Full reliance is placed on the sincerity of Mr. Brown in his
written reports of the data, but no reliance is placed on his
interpretations. His orally reported data have been partially
discounted as being less reliable than his written records.
There is no evidence of conscious deceit.
Most of the interviews have taken place at the Townsend Brown
Foundation, 306 North Vermont Avenue, Hollywood, California.
4 ) The Basic Effect
At Zanesville, Ohio in 1927, Mr. Brown at this then Laboratory
of Research Devices, performed experiments on force acting upon
charged conductors. The apparatus took two forms, as described
in a manuscript "Tapping Cosmic Energy", dated 28 October 1927,
which was submitted at that time to the Physical Review
for publication but which was rejected. Apparently no basic
experiments have been performed since 1928.
4.1 ) In one form the conductors were two independent
spheres suspended close but not touching by means of fine wires
which also served as electrical leads. The spheres were of lead,
and weighed 14,740 gr each. The wires supporting them hung from
the ceiling 45 cm apart. When the spheres were oppositely
charged to a potential difference of 125 kv they moved toward
each other slightly in response to their electrostatic
attraction. The negative sphere was displaced allegedly about
twice as far as the positive sphere. The inequality of the
displacement is the essence of the phenomenon.
4.2 ) In the other form of the basic experiments the
condenser was formed by a block of insulating material 3-1/3 c
2-1/2 x 15", to the ends of which electrodes were attached.
Whether the condenser was suspended or whether it was mounted at
the end of a centrally pivoted horizontal beam, it showed a
force causing the arm to rotate, usually in such a direction
that the positive electrode moved forward. The force is claimed
to have been proportional to the mass of the block, although no
data are given in the manuscript and none are available. The
name "gravitor" is used for this apparatus.
The condenser dielectric was either homogenous or laminated. In
one laminated version, 10,000 sheets of 0.001" lead foil were
separated by an equal number of 3 x 3" sheets of 0.0008"
cellulose acetate in a formica box, to make 10,000 condensers in
a series arrangement weighing 10 kg. The acceleration exhibited
by this device when suitably charged is claimed to have been 2.5
cm per sec2 when the electrodes were charged, evidencing a force
of 25,000 dynes or about 25 gm; the homogenous type is stated by
Mr. Brown to have had comparable thrust. The force was usually
toward the positive electrode, although in some units it was
toward the negative; this effect was attributed to a
"polarization". The force is stated to have been stronger at
some time that others, correlating with the position of the sun
and moon. Data are displayed which relate to the angular speed
of a pivoted gravitor, and which show half a dozen large
fluctuation in a day; some of these can perhaps be construed a
being dips simultaneous with the rise or setting of the sun or
On 17 June 1952 the undersigned accompanied (censored) of the
Townsend Brown Foundation to the Glendale plant of the Bendix
Aviation Corporation. An experiment of the type described in
paragraph 4.1 had been set up by Bendix personnel. (Sentence
censored). The observations in Table 1
were recorded by the undersigned.
Table 1: Observations by the Writer on
17 June 1952 Experiment at Bendix Aviation Corp., Glendale, CA
From a wooden scaffold a cylindrical brass weight 8" long and
4" in diameter was hung with axis vertical by means of about 5'
of #14 Cu wire (0.06408" diameter). A similar wire about 6" away
suspended a similar weight level with the first. By means of the
wires the weights were electrified to a measured voltage; the
corona current was not measured. An optical projection system
provided on the wall an image on a scale and insulated pointers
attached to the weights. The weights were either hollow (5 lb)
or solid (30 lb). In certain of the observations the weights
were linked by an insulating bar.
4.3 ) An examination of the data makes it clear that the
inequality in deflections reported by Brown is still present
here, though only to the extent of 1-30% (tests 2-4) and less
than that under conditions where the corona discharge is reduced
(of tests 1 & 5).
When cigarette smoke was drifted into the vicinity of the wires
it showed the existence of a wind. Just above the weights the
wind blew from negative top positive, while higher on the wires
the wind was opposite.
5 ) The Brown Electrometer
In a small thermostated room in the second basement of the
Banks-Huntley Building (address censored), Los Angeles, there is
a continuous experiment which shows an extension of the basic
phenomenon. This apparatus is the Brown Electrometer. The device
is housed in an oil-filled tank of half-inch steel, perhaps two
feet in diameter and a foot deep, with a half-inch steel cover.
In the tank there is suspended a horizontal disc 12" in diameter
and one inch thick, made of six segments of pine wood
alternating with six of marble around an 8" hub of Bakelite. The
sections are separated by 1 x 2 x 1/16-inch copper strips.
Alternate strips are grounded to a central metal stem supporting
the disc which is attached to a music-wire suspension 0.018" in
diameter and 6-3/8" long; the remaining copper strips connect to
a central iron tit at the center of the hub, which dips into a
cup of mercury. The disc is surrounded by a grounded metal
shield which is attached to the tank and which stands at a
clearance of 1/2-inch above and below the disc and 1-1/2 inch
radially; this shield is centrally perforated at the top and
bottom for the suspension and the tit.
A sequencing device in a cabinet below the tank causes a
half-wave smoothed DC power supply to impress upon the mercury
cup a well-regulated negative potential of 11.1 kv resulting in
a leakage current of 0.45 ma. The sequencer leaves the voltage
on for 30 seconds, causing the disc to rotate several degrees.
After 30 seconds, a short spark is made to jump from a point
projecting about 9 inches horizontally in air from the stem of
the disc, and perforates a rice-paper strip wound on a drum
which rotates once each day. The voltage is then turned off and
the disc returns to its relaxed position, which is recorded by a
separate spark. The cycle repeats every three minutes. The
apparatus appears to be well engineered and extremely reliable.
5.1 ) Mr. Brown has given attention for several years to
the variations in the deflections of this electrometer. On the
spark record the rest position is 50 units and the deflected
reading is about 80. Until about a year and a half ago the
deflected position was about 20, but suffered a gradual change
to its present value at that time. Aside from the reversal of
sign, the deflection undergoes fluctuations whose character is
less dramatic but still large, e.g., 10-50%. The cause of these
fluctuations has been sought by Mr. Brown by seeking for
correlations between them and various outside influences, such
as magnetic storms, temperature, pressure, the time of the year,
civil time, lunar time, sidereal time, and the Dow-Jones
average. Only the last five correlations are claimed. Tabulated
data are available only for the year 1937. Hourly recordings
throughout that year are tabulated, both for civil and lunar
time. The writer has seen no convincing evidence that the stock
market and the automatic recorder are correlated, although it
has been claimed that a chi-square analysis has been performed,
with positive results.
5.2 ) The undersigned has conducted a partial statistical
analysis of the data from the automatic recorder for 1937 as
5.2.1 ) Use was made of a table furnished by Mr. Brown
entitled "Hourly Readings of Instrument, 1937" and particular
attention was directed toward the readings of 0100, one hour
after midnight, on each day of the first half of 1937. During
the first half of January, for instance, the readings on days at
this hour were 26.20, 25.90, 26.20, 26.10, 26.80, 25.60, 25.40,
26.00, 26.00, 25.60, 26.60 and 25.90. The mean is 26.05, the
standard deviation 0.55 and the probable error of the mean 0.15.
Table 2 gives the half-monthly
means and probable errors for time 0100 and for time 1300. This
table shows that there was a strong and real variation during
the spring of 1937; but the great similarity between the
readings at 0100 and 1300 implies that there is no conspicuous
diurnal effect. (See also paragraph 5.2.2)
5.2.2 ) A direct search was undertaken by the writer to
see whether the data for the first half of 1937 reveal any solar
diurnal period. Attention was confined to the first day in each
half-month. On these twelve selected days the readings at time
0200, for example, were 26.20, 26.00, 27.90, 29.00, 28.10,
29.80, 30.40, 29.80, 28.90, 27.50, 27.40, and 28.10. The mean
was 28.29 and the probable error of the mean 0.46. The probable
errors, as far as calculated, were all in this neighborhood. Table 3 gives the means calculated as
in the above example. It is seen that there is no strong diurnal
periodicity, with the dubious exception of a peak at 1100 hours;
this result confirms that of paragraph 5.2.1.
5.2.3 ) The undersigned has scrutinized a table of data
captioned "1937 Readings of Instrument with Lunar Hour Angle" in
a search for evidence of a variation of the readings with the
position of the moon, as rising, setting, etc. The form taken by
this search was as follows. For the calendar month January the
average deflection for each lunar hour was noted; e.g., at lunar
hour angle 1300, when the moon was one hour past the meridian,
the readings for January averaged 26.74. At the same lunar angle
in the succeeding calendar months the average readings were
28,60, 30.18, 29.50, 28.02, and 28.52. The average of those six
averages gives 28.59 as the mean of all observations taken at
lunar hour angle 1300 during the first half of 1937. This
process was repeated for each of the 24 hours. The means are
listed in table 4. Their constance is most striking; they show
no dependence upon lunar hour.
Table 2: Half-month mean deflections at
two times a day, during the first half of 1937
Table 3: Hourly deflections averaged
over the first day of the half-month in January-June 1937. The
Probable errors are above 0.46
Table 4: The deflection of the
instrument as averaged over the first half of 1937 for each
lunar hour angle
5.3 ) By the foregoing sketchy analysis of the data from
the Brown Electrometer, the following conclusion are implied:
(a) Through the weeks and months the data drift significantly;
(b) There is no evidence of a solar diurnal period synchronizing
with the sun, as claimed by Brown;
(c) There is no evidence of a lunar diurnal period synchronizing
with the moon, as claimed by Brown. On the contrary, the data
appear most emphatically to deny such a period.
5.4 ) Two related experiments have been described by Mr.
Brown which throw light on the reason for the deflections if not
the reason for their fluctuations.
5.4.1 ) In one, the sector disc was held stationary while
a grounded metal disc was suspended horizontally over it in the
oil, by a torsion fiber. When the sector disc was charged, a
motion of the adjacent oil was noted by Mr. Brown. The suspended
disc deflected, but not when a glass plate was interposed
between the sector disc and the suspended disc.
5.4.2. ) The other related experiment concerns an
oil-immersed annulus described by Mr. Brown as 2 feet in
diameter and 6 inches thick, made of lead oxide (PbO) and
paraffin in eight segments separated by 8 copper strips
alternately charged positive and negative. By means of a probe
the potential distribution on the PbO and paraffin surfaces was
explored and it was noted that whereas the gradient was uniform
across the edges of the lead oxide sectors, the gradient on the
paraffin was steepest near the negative electrodes. This
observation means that the steepest gradients are all in the
same direction, e.g., clockwise. It was also noted that the
application of high voltage to the electrodes of this disc
caused a lively motion of the oil.
6 ) The Model Flying Saucers
At the quarters of the Townsend Brown Foundation a room is
devoted to demonstrating a new method of propulsion (para. 2.2)
which is claimed to depend upon the electrogravitic effect.
6.1 ) At the center of the room there stands a tall iron
post carrying at its top a transverse Plexiglass beam about 6
feet long, which is freely pivoted at its center to rotate in a
horizontal plane. From each end of the beam hangs a model flying
saucer, Figure 1, of which the metal
part is composed of two spun aluminum discs 12" in diameter,
between which is bolted an 18" Plexiglass disc. The periphery of
this disc is indented with nearly semi-circular scallops. The
fingers remaining between the scallops constitute standoff
insulators supporting a bare 0.005-inch stainless steel wire AA
at a distance of 3 inches from the aluminum spinnings and
insulate it from the aluminum. This will be called the outboard
wire in the following discussions. It embraces perhaps
one-quarter of the circumference. A similar wire, BB, is
supported at the opposite quadrant, but is connected to the
aluminum spinnings; it appears to be without effect. Each of
these two identical assemblies is suspended by an insulating
string from one end of the Plexiglass beam. By rubber-covered
leads, shown in Figure 3, from the
beam spinnings can be given one polarity while the outboard
wires are given the opposite polarity. The leads are energized
through rotating contacts at the pivot. If the beam rotates
counter-clockwise as seen from above, the discs revolve about
the post, in such a way that the outboard wire is at the leading
A full-wave rectified unsmoothed high-tension supply in an
adjacent room provides voltage for the demonstration. The
voltage is controlled by a Variac. An examination showed no
6.2 ) When the power was turned on, the discs became
negative and the outboard wires positive. Corona discharge could
be heard as a gentle hiss or crackling. It could be seen in the
dark, particularly around the outboard wire. As soon as the
potential was applied, the discs and the beam began to revolve,
the leading edge being at the outboard wires.
6.2.1 ) The thrust was roughly measured by a method in
which the rotation was prevented, as shown in Figure 3. The weight on the pan was
adjusted until the indicated angle was 45 degrees, at which time
the thrust was known to be equal to the force in the vertical
thread. This method of thrust measurement is valid if the pivot
is frictionless; actually, the friction seems to introduce an
uncertainty of about 0.5 gm in the thrust data. The procedure
described provides a value of the thrust of the two discs
Figure 3: The method involving three
threads A, B & C as used to measure the thrust of a
demonstration flying saucer
6.2.2 ) The data shown in Table 5
were gathered by the writer on 6 June with the collaboration of
Mr. Brown and (censored), an assistant. Column 1 gives the
setting of the Variac in the supply circuit for the high-tension
transformers. The corresponding voltages, with respect to ground
at the positive and negative outpost cables from the rectifier
are given in columns 2 & 3; at each polarity the voltage was
measured with the aid of a 0-2 milliameter and a high series
resistor. The current in each lead, recorded in Columns 4 &
5, was measured by the same milliameter when appropriately
reconnected; it may be resumed that most of the current
represented corona discharge between the discs and the outboard
wires. Column 6 gives the measured thrust. In Column 7 is
recorded the time taken for the discs to revolve once completely
when free to do so, after they had had time to reach their full
speed. The entires in Column 8 give the thrust to be expected on
the basis that it is caused by the electric wind (see para. 7
below). The last Column shows the efficiency of the propulsion,
calculated (with due heed to the conversion of units) from the
ratio of the electrical power input (Col. 2 x Col. 4 plus Col. 3
x Col. 5) to the mechanical power output (Col. 6 x Velocity).
The data in Columns 2 & 3 were gathered first, then those in
Columns 4, 5 & 6, and finally in Columns 2 & 3 again.
The repeat measurements of the voltages nearly always agreed
with the first measurements to within 0.75 kv or less. The
average is recorded in Columns 2 & 3.
Table 5: The results of measurements on
the demonstration apparatus for flying saucers
6.2.3 ) Three additional observations were made. In one,
slips of thin cardboard were cut, folded double and so hung on
the outboard wire of each disc as to cover practically its
entire length. Their purpose was to confine the positive ions to
the region of the wire and so reduce the corona. It was observed
that the thrust, originally about 7 gr as shown in Table 5, was reduced to about 1 gr.
6.2.4 ) In the second additional observation the polarity
of the discs was made positive while the wire was negative. The
thrust was then 3 / 1 gr (in the same direction as before)
instead of about 7, the Variac setting being 40, 45 and 50.
6.2.5 ) The third additional observation was that when the disc
was charged and immobilized as in Figure 2,
a brisk wind was developed. The origin of the wind was
apparently near the outboard wire. Its direction was from the
wire toward and past the aluminum spinning as shown by cigarette
7 ) Possible Theories of the Townsend Brown Phenomena
7.1 ) Mr. Brown has no theory for the effects that he has
demonstrated or claimed. He has advanced an empirical
explanation for the magnitude of the force on a suspended
condenser, but has not advanced supporting evidence. While
granting the existence of the electric wind, Mr. Brown seems
nevertheless to believe that some new force is operating which
he describes as electrogravitic. This force is variable with
time as shown by the Brown Electrometer. The origin of the force
is of smaller interest to Mr. Brown than is the cause of its
variations. No cause for the variations has been proposed, other
than the alleged correlations with astronomical bodies.
7.2 ) It might be suggested that electrostriction is
significant in Mr. Brown's effects. This suggestion is untenable
because (1) the deflection of a suspended condenser (Table A
[sic], Tests 6 & 7) is far greater than any conceivable
electrostrictive expansion or contraction; (2) electrostriction
would cause no acceleration of the center of the center (sic) of
gravity; and (3) the steady rotation of the flying saucer models
cannot be caused by electrostriction, which is a dynamic effect
only while the voltage is changing, and not in the steady state.
7.3 ) The force exerted upon a wire carrying a current in
a magnetic field could conceivably contribute to the phenomenon.
This hypothesis appears tenable at first sight, since the earth
always provides the field, and with the corona discharge there
is always associated a current. It is easily disproved, however,
by (1) the sign of the deflections shown in Table 1 was unchanged in test 3 when
the polarity, and thereby the direction of any magnetic effect,
was reversed; (2) the direction of the thrust by the model
flying saucer was unchanged by changed polarity; (3) the equal
and opposite corona currents in the leads to the saucers should
result in equal and opposite magnetic force, and zero net
thrust; and (4) there is no report that the Brown Electrometer
shows normal deflections during magnetic storms.
7.4 ) The writer proposes that the origin of the observed
phenomena is the electric wind. The effects described above all
relate to the forces upon electrically charged bodies or upon
adjacent insulating material. Such forces have long been the
subject of observation and mathematical theory, and are
discussed accordingly in such works as W.R. Smythe's "Static and
Dynamic Electricity". It is commonly held by physicists that no
practical discrepancies exist between theory and experiment in
static electricity, if the experiments are conducted in the
manner contemplated in the theory. It must be emphasized in this
connection that the theory of electrostatics is written around
the assumption that the insulators in the apparatus are perfect
non-conductors, and that the fluid (e.g., air or oil) in which
the apparatus is immersed is also a perfect non-conductor. For
an experiment performed in a non-conducting medium the verdict
of electrostatics is quite plain: the Townsend Brown effects
should not exist.
The experiments by Mr. Brown, including all those described in
paras. 4, 5 and 6 were carried out in a partially conducting
medium, i.e., air or oil. In such a medium the simple laws of
electrostatics are not applicable. It is true that in the air or
oil the conduction of electricity is usually so slight that the
departures from the laws of electrostatics are quite negligible.
When, however, large voltage gradients are involved, the density
of ions is very greatly increased for well-known reasons, and
the conductivity of the medium increases commensurately. If the
voltage is high enough the passage of the ions through the air
causes the crackling sound and bluish glow characteristic of the
corona discharge. In the presence of such a discharge, the laws
of electrostatics are not expected to hold. We will show that,
so far from being forbidden, the Townsend Brown effects are
actually predicted in the presence of a corona discharge.
7.5 ) We will adapt the treatment cited in (name
censored) ("Conduction of Electricity Through Gases, 3rd ed.,
Vol. II, 1933, page 544) in considering the force on a charged
conductor brought about by the ionic conductivity in a gas.
Consider a positively charged fine wire in air, parallel to a
negative conductor. The well known concentration of the electric
field near the fine wire will then produce a corona discharge;
positive ions are generated at the wire. The wire being
positively charged repels the ions. The force on the wire is
roughly equal and opposite to that on the ions. If E is the
field intensity and u the space charge density, the
force on the ion is:
Where the integration covers the region occupied by the space
charge. Here if we ignore the non-parallelism of the E vectors,
we may write roughly
It is known that the velocity W of an ion in a fluid is related
to the electric field in which it moves through the relation
W = kE
Where K is ionic mobility.
For a positive ion in air at atmospheric pressure its value
will be taken as 1.32 (cm/sec)/(volt/cm), while for the negative
ion the value will be taken as 2.51 (cm/sec)/(volt/cm).
is the current,
Where d is the distance between the conductors. This force on
the ions is communicated to the air by molecular collisions, and
creates what is known as the electric wind; it is observed in
the neighborhood of sharp points as well as fine wires. The
thrust may be regarded as the force of reaction which results
from the generation of the wind; the thrust is in the direction
opposite to the wind, as in a jet airplane.
7.5.1 ) It is known that the mobility of a negative ion
is greater than that of a positive, the ratio being about 1.9 at
atmospheric pressure. The force on a positive conductor should
therefore be, and according to (censored) and (censored) is
observed to be, greater than that on a negative conductor.
7.5.2 ) Further confirmation of the above theory is found
in the variation of F with pressure. Theory predicts that since
k is well known to vary as the inverse square of the pressure
the force should vary directly as the square root of the
pressure; this prediction agrees with the early results of
Arrhenius, Wied. Ann. 63: 305 (1897).
7.5.3 ) The efficiency of the electric wind as a
propulsive agent is easily calculated. The power output is the
product ViD/k of the thrust and the vehicle velocity V; while
the power input is iv is the potential difference. The
efficiency is then given by
At a given vehicle velocity the efficiency can be improved by
increasing the electrode spacing D or by decreasing the voltage
necessary to sustain the corona.
7.6 ) Physically the theory of the electric wind is a
shown in Figure 4, which shows a
section view of two parallel charged wires in air. Positive ions
are formed at the positive wire and surround it, but the
attraction of the negative wire draws the positive cloud so that
it lies primarily to the left of the positive wire in Figure 4. The slow mobility of the
positive leads to a considerable positive charge in this region,
which is repelled by the positive wire and which repels it.
Negative ions are generated at the negative wire, and form an
unsymmetrical cloud about it. Since their mobility is fairly
fast, they are less crowded than the positives. The force on the
positive wire is made of two parts; attraction to the left by
the negative cloud (e.g., 20 gr) and repulsion to the left by
the ion cloud (e.g., 10 gr). the force on the negative wire is:
attraction to the right by the positive wire (e.g., 20 gr) and
repulsion to the left by the ion cloud (e.g., 5 gr). The net
force on the positive wire would in this hypothetical case be 10
gr to the left, while that on the negative wire would be 15 gr
to the right. The difference, 5 gr to the right would be
directly observed if the weights hanging from the wires were
connected by a rigid insulating link.
Figure 4: The ion clouds near two parallel
oppositely charged wires in air
8 ) Test of the Electronic Wind Hypothesis
It is believed by the writer that the electric wind in air and
in oil leads to all the well-documented results described
herein. These results will be summarized and compared with the
8.1 ) (See para. 4.1) The positively charged sphere in
the Zanesville experiments moved less that the negatively
charged weight. This is as expected because the electrostatic
force on the positive weight was partly balanced by the kinetic
reaction from the electric wind which presumable existed (para.
7.5 & 7.6) because of the corona discharge at the wires.
8.2 ) (See para. 4.3) The inequality in the motion of the
weights in the Bendix experiment was less noticeable, thus
discouraging corona and reducing the current. The observed
electric wind was caused by corona at the wires, which was more
or less strong according as the wires were rough or dusty in
8.3 ) (See para. 4.2) the deflection of a suspended
condenser when charged was claimed. This is as expected if the
leads or terminals sustain a corona. The "polarization" effect
would seem to be evidence that one or the other lead or terminal
had rough areas, sharp corners, or dust resulting in corona at
one end to the other, and a thrust in one or the other
direction. If the terminals have the same geometry, the force
will, because of the lesser mobility of the positive ions, be
directed toward the positive terminal; if the negative terminal
is much rougher or sharper, it will be the seat of a corona of
which the force might outweigh that at the positive terminal.
The observed fluctuations in the effect are to be expected as
the result of changing dust on the wires or terminals, which
changes the corona current; the deflections in the Bendix
experiment were somewhat irregular.
8.4 ) (See para. 5) The disc of the Brown Electrometer
shown in Figure 5 develops a torque in response to an
applied voltage. The field in the oil is strongest in the
regions A.B.V.D. where the oil is in contact with the electrodes
separating the pine sectors from the marble sectors M. Positive
ions formed at the electrodes as in a corona discharge will
produce positive space charges at the regions A. & D. while
regions B. & C. will have negative space charges. In the oil
these will cause effects analogous to the electric wind in air.
If the fields A.B.C.D. are equally strong the ion densities will
be such as to give no net torque to the disc. We will assume, on
the other hand, that the surface of the pine sectors allows a
leakage current; this will have the effect of reducing the field
in the regions A. & B. and will leave the regions C. &
D. dominant in the production of force. The positive ions at D.
will tend to rotate the disc clockwise while the negatives will
tend to rotate is counterclockwise. Since their mobilities
differ there will be a net torque in one or the other direction,
as observed. If the chemistry of the oil or the insulating
sectors changes with time the surface conductivity of the marble
could become larger than that of the pine, causing a reversal of
the torque. Aside from surface conductivity, other attributes
may distinguish the marble from the pine. For example, a net
torque might be caused by the influence of different dielectric
constants from the field distribution, or if alternate sectors
all project slightly from the electrodes.
Figure 5: Electrometer
A net torque might result from or be influenced by the surface
conductivity of the pine and marble. If one is more conductive
than the other, the field pattern will be made unsymmetrical
(See para. 5.4.2) and the torque will be altered. It is possible
that the slow fluctuations in the Brown Electrometer are caused
by changes in the surfaces. The faster fluctuations may be
caused by particles suspended in the oil, which may temporarily
rest at such points as A. or B., changing the corona there
8.5 ) The experiment reported in para. 5.4.1 would seem
to corroborate the idea that a swirling of the oil is essential
to the operation of the electrometer. This idea is in accord
with the electric wind theory, whether or not the discussion in
para. 8.4 is valid in detail.
8.6 ) It is claimed that the deflections of the
electrometer show certain periodicities. If these are real they
show the inadequacy of the electric wind hypothesis. Their
reality is assessed in the crude and preliminary analysis
contained in para. 5.2 and summarized in para. 5.3. There is no
manifest solar diurnal period and no manifest lunar diurnal
period. No test has been applied for the claimed sidereal
diurnal period or for the correlation with the value of
financial securities. Neither has it been possible to determine
whether the slow changes in deflection (Table 2) are repeated
annually as claimed; however, the fact that the deflections once
annually permanently reversed direction would seem to discourage
this claim. This reversal is attributable to changed surface
conductivity of the sectors of the disc. It is concluded that
the short study, para. 5 of the reading of the Brown
Electrometer does not weaken the hypothesis that all the
(censored: probably reads, "phenomena reported by Mr. Brown")
are the result of electric wind.
8.7 ) The thrust observed in the experiment on the model
flying saucers is given in Table 5,
para 6.2.2. The thrust to be expected on the electric wind
hypothesis is tabulated as calculated from the equations, para.
7.5. This agreement is satisfactory in order of magnitude.
Actually the calculated thrust is about twice the observed. The
discrepancy may be due to the approximations made in the
mathematics of para. 7.5.
8.8 ) (See para 6.2.3) A partial shielding of the
insulated wire on a model flying disc partially eliminated the
thrust, presumable because the current was probably thereby
reduced. The shielding also had the effect of making the
approximations in para.7.5 still worse. Qualitatively, then, the
electric wind hypothesis is again verified.
8.8 ) (See para. 6.2.4) With polarity reversed so that
the insulating wire was negative, the thrust fell form about 7.5
gr to about 3 gr, the ratio being 2.5. Unfortunately the current
was not measured. Assuming it to have been unchanged in
magnitude by the reversal, we would predict, from the result,
that the mobility of the negative ions is 2.5 times greater than
that of the positives. This is in fair agreement with the known
factor of 1.9.
8.10 ) (See para. 6.2.5) A wind was observed to be
generated near the outboard wire of the model flying saucers.
This is as expected on the hypothesis under evaluation. The
thrust experienced by the disc may be regarded as the jet
reaction caused by the formation of this wind.
9 ) Conclusion
9.1 ) All the well documented effects disclosed by Mr.
Brown appear to be explained as caused by ionization of the air
or oil in which the apparatus is immersed.
9.2 ) Mr. Brown claims that a gravitational anomaly
exists in the neighborhood of a charged condenser. This effect
has not been well documented by Mr. Brown, nor has the writer
undertaken to refute it.
9.3 ) The thrust developed by the model flying saucers
has its origin in the well-known phenomenon of the electric
9.4 ) The efficiency of propulsion by the electric wind,
as exemplified by the model flying saucers is of the order of
1.4 percent. At speeds higher than 2.3 ft/sec attained by the
model flying saucer the efficiency might be substantially
improved. It is presumable less at high altitudes and zero in a
vacuum.* If the efficiency of conversion of fuel energy into
electrical energy is 21 percent the overall efficiency of
propulsion of the model flying saucers by the electric wind is
0.3 percent. This compares with about 25 percent for a
propeller-driven airplane and about 15 percent for a jet
End of Report
Classification of this report cancelled by authority of Office
of Naval Research Ltr. 2340, by Capt. Dewey Brown, CI Div.,
Dir., Special Investigations, HQ USAF, Wash. DC., 1 October
* N.B.: The "presumably less at high altitudes and zero in
vacuum" hypothesis was later disproven by Brown during
experiments in France, where it was shown that in fact the
efficiency of his model flying saucers actually increased when
flown in high vacuum. ~ Ed.
N.B.: The remainder of the report is comprised of miscellaneous
military letters that do not contain any pertinent information.