Chapter 5
Simultaneous strains: the 'surface of action'
There have been many reports by families of several metal objects being found bent at the same time. Of course it is difficult to observe two widely separated objects actually bending at the same moment, since the eyes cannot subtend a wide field of view; but occasionally there are good reasons to suppose that two or more bendings may have taken place simultaneously. We are forced to pose the questions: Did these involve simultaneous strain pulses, and if so, how extensive was the region of action? Were any strains experienced elsewhere, and where was the subject at the time? Does his location affect the strains,: and can he direct his 'powers' at several objects at the same moment? Is there conceivably a terrestrial effect a preferred orientation for signals?
A physicist might suppose that some sort of wave-front issues from the subject, like ripples from a stone dropped in water. If this were so, objects equidistant from him would receive signals at the same instant. Does this occur?
We have investigated these questions using the strain gauge equipment described in the previous chapter. Synchronized chart-records and independent strain gauges, bridges, amplifiers and batteries were used.
The first studies with two, and later with three, independent strain gauges were undertaken in thirteen sessions with Nicholas Williams during the first half of 1976. At first he simply attempted to influence both chart-records while the two metal specimens were about a foot apart; his own position was not fixed during the session. We found that signals did appear on both chart-records, sometimes simultaneously, and therefore it was possible to proceed to more controlled experiments.
Since it was necessary that Nicholas should be in one location during the entire session, a suitable occupation had to be found for him. He was already an experienced builder of model aircraft and ships, using commercial assembly kits, glue and paint. He was asked to do this at a work surface in the open-plan ground-floor living-area of his house; individual latchkeys, each containing its own strain gauge, were arranged in suitable locations, suspended from their own electrical connections, so as to minimize mechanical coupling between them. A plan of the area is shown in Picture 5.1; the circle S represents Nicholas's location, but that of the observer (myself) is not shown. It varied between the lounge sofa, the kitchen and even the stairs; and sometimes I would have to move to the apparatus to make adjustments. As far as possible, other observers were excluded from these sessions, but on some few occasions the presence of Nicholas's father, Dr Terry Williams, Dr Miller and one or two family friends was recorded.
Figure 5.1 Plan of ground floor of Nicholas Williams's house, in which sessions with two and three strain gauges were conducted.
S. the subject, Nicholas Williams; 1EH,2EH, two strain gauges in the equidistant horizontal configuration; 1RH,2RH, two strain gauges in the radial horizontal configuration; 1OH,2OH, two strain gauges in the opposite horizontal configuration; RHV, two strain gauges in the vertical configuration, or in the radial-horizontal vertical configuration (with RH).
L, lounge; D, dining area; K, kitchen.
My own position was on the lounge sofa, in the kitchen, or occasionally watching from the stairs.
The routine was as follows: I would arrive around noon and set up the equipment. It was then allowed to run quietly in the empty house for as long as possible, with Nicholas out of the house for as much as two hours. When we returned, the chart-records were found to be free from artefacts. The afternoons were used for exposures of about two hours each. On several afternoons the suspended latchkeys spontaneously bent to and fro, and ultimately fractured.
The latchkeys were suspended in various configurations with respect to Nicholas, as described in the caption to Figure 5.1. There were four configurations with two latchkeys: equidistant horizontal, signified by the caption EH; radial horizontal, signified by RH (here the two keys were on a horizontal line stretching radially from Nicholas); opposite horizontal, OH (in which they were horizontal and at directly opposite sides of Nicholas); and vertical, V, in which one latchkey was directly above the other, at the position shown.
When three latchkeys were used they were mounted in the radial-horizontal-vertical configuration, RHV. The three keys defined a vertical plane stretching radially outward from Nicholas.
For the most rigorous experimentation it would be necessary to change from one configuration to another in a random way, possibly after every signal event or even at randomly selected times. Such procedures would cause considerable disturbance in the living-area and of course could not be concealed from Nicholas. I decided that he should know what the configurations were; no attempt would be made to screen the sensors from view; but the decisions as to when to change the configurations would be made by me, usually after a small number of events. This would not be a completely random procedure, but regularity in the changes was carefully avoided. In later sessions I made efforts to induce Nicholas to learn to induce synchronous signals in each configuration; when he achieved success, the configuration would be changed; ultimately a change would take place after every three events. These procedures were followed because of my increasing certainty that the configurations were matters to which Nicholas would probably react differently at each session; at first he did not express much interest in what configuration was being offered; but in the end he was motivated to try as hard as he could to produce the maximum 'control' over the sensors in each configuration. This 'control' was understood to be the production of synchronized signals on more than one sensor.
Table 5.1 Classification of events in Nicholas Williams sessions A-G
Configurations: EH, equidistant horizontal; RH, radial horizontal; V, vertical; RHV three sensors, radial horizontal and vertical with two vertical sensors closest to subject; 30°, inclination of subject to line joining two horizontal sensors (90° would represent EH); Up, subject upstairs, sensors downstairs; OH, opposite horizontal; RHS, radial horizontal with one sensor screened; OHS, opposite horizontal with perspex screen over fine sensor.
Synchronization classifications: U1, 2, signal unique on 1 or 2; S. synchronous; NS, non-synchronous; SS, signals on one sensor suppressed (small magnitude), but still synchronous.
Session Date (1976) Designation Configuration Synchronization Range (mV) Remarks
A 9 Mar 1 ~EH U1 0.1
2 ~EH NS 0.1
3 ~EH S 0.1
4 ~EH S 0.1
B 18 Mar 1 30° S 0.1
2 30° NS 0.1
3 30° NS 0.1
4 30° S 0.1
5 30° U2 0.1
6 30° S 0.1
7 30° NS 0.1
8 Up S 0.1
8a 30° U2 0.1
9 30° S 0.1
10 30° S 0.1
11 EH NS 0.1
12 30° NS 0.1
13 RH S 0.1
14 30° U2 0.1
C 23 Mar 1 RH S 0.1
2 RH S 0.1
3 RH S 0.1
4 RHS NS 0.1 Screening by kitchen foil
5 RHS NS 0.1 Screening by kitchen foil
6 EH NS 0.1
7 EH NS 1
8 EH NS 1
9 EH NS 1
10 EH NS 1
11 EH NS 1
C1 23 Mar 1 EH S 0.1
2 EH S 0.1
3 EH NS 0.1
4 RH S 0.1
5 RH S 0.1
6 RH NS 0.1
D 9 Apl 1 RH S 0.1
2 RHS NS 0.1 Glass screen
3 RHS NS 0.1 Glass screen
4 RHS U1 0.1 Brass screen on key 1 which bent to 70°
5 RHS NS l Steel screen on key 1 which bent to 135°
6 RHS NS 1 Steel screen on key 2, which bent to 60°
7 RH NS 1 Key 1 fractures, key 2 bent at 60°
8 RH NS 1 Key 2 bent to 80°
9 EH NS 1 Key 2 bent to 85°
10 EH NS 10 Key 2 partially fractured
11 EH U 10 Key 2 fractured
E 15 Apl 1 OH U1 0.1 Fracture of table spoon with sensor
2 OHS NS 0.1 Perspex screen on eutectic alloy specimen
3 OHS NS 0.1 ditto
4 OHS NS 0.1 ditto
5 OH NS 0.1
6 0H S 0.1
7 OH U1 0.1
8 OH S 0.1
9 OH S 0.1
10 0H U1 0.1 Intended key 1 only
11 OH U2 0.1 Intended key 1 only
1 23 Apl 1 V S 0.1 Unwitnessed alleged paranormal shooting of piece of metal from tube
2 V S 0.1
3 V S 0.1
4 V SS 0.1 25° bend on key 1 Failure of chart pen
5 V SS 0.1 45° bend on key 1, 25° bend on key 2
6 V NS 0.1 Some misalignment of configuration; 35° bend on key 1, 25° on key 2
7 V SS 0.1 Some misalignment of configuration
8 V SS 0.1 10° bend on key 1; 45° on key 2
9 V SS 0.1 10° bend on key 1; 45° on key 2
10 V S 0.1
11 V NS 0.1 Intended delay, 135° bend on key l; 45° on key 2
12 V S 0.1 Key 1 fractured
13 EH NS 0.1
14 EH SS 0.1 Some misalignment of configuration
15 EH S 0.1
16 V SS 0.1
17 V NS 0.1
18 V S 0.1
19 EH SS 0.1
20 EH S 0.1
20aEH S 0.1
21 V S 0.1
22 EH S 0.1
23 RH S 0.1
24 RH S 0.1
G 30 Apl 1 RHV S 0.1
2 RHV S 0.1
3 RHV S 0.1
4 RHV S 0.1
5 RHV S 0.1
6 RHV S 0.1
H 27 May 1 Single 0.1 All viewed on television monitor
2 Single 0.1
3 Single 0.l
4 Single 0.1
5 Single 0.1

In some events there are signals on one sensor only. In some, both traces display signals, but one is obviously in advance of the other. But in some the signals appear, within the accuracy of the equipment (0.2 see), to be recorded at identical times. The signals are not of the same magnitude, or even in the same sense, but their time structures are very similar; they are, basically, synchronous signals.
The classification of the data from all these sessions into synchronous and non-synchronous signals is clearly a valuable exercise. Usually it is as simple as has appeared from the above example, but there are a few signals in which the classification is doubtful, and these have not been included. Table 5.1 lists the classification, which can be summarized as indicated.
ConfigurationSynchronous signals Non-synchronous and unique signals
Radial horizontal 9 3
Vertical 6 3
Radial-horizontal-vertical 6 0
Opposite horizontal 3 1
Equidistant horizontal 4 (+ 6 in session F) 13

It appears that synchronism is most characteristic of the radial horizontal, the vertical and the radial-horizontal-vertical configurations. But it is not characteristic of the equidistant horizontal configuration, apart from the very successful 'learning' of session F. If the situation were reversed, one might seriously consider the possibility of a circular or spherical wave-front proceeding outwards from the subject. The very existence of large numbers of synchronous signals makes it necessary to consider what might be causing them. The experiments establish that essentially simultaneous strainings take place, and that most commonly they take place on a vertical plane passing through the location of the subject.
It could be that synchronous paranormal action was taking place elsewhere as well, but we had at that time no evidence for this; the minimum hypothesis is that it took place on a surface which is not necessarily planar but is sufficiently so for three widely separated points on it to lie on a vertical plane. Let us propose the following physical model: that there can exist, in the neighbourhood of a subject, a 'surface of action', at points on which strain occurs on objects. We do not know that all points on the surface are so affected, we have information about only three (and more usually, only two; but in the further experiments with Stephen North with as many as six strain gauges, the results described below are essentially similar); nevertheless, strain throughout the surface of action is the simplest assumption. We can then discuss the phenomenon in terms of the extent, the configuration and the movement of the surface of action. This surface might plausibly be regarded as a sort of invisible extension of the human body, perhaps even of the human arm or the haptic system.
The surface is considered to move slowly about the room, presumably under the influence of the unconscious mind of the subject. When one strain gauge displays a signal in advance of the other, it is because the surface has moved from one to the other. Although we do not know that it has moved directly, or at uniform speed, we might estimate from the data that if we did assume these things, the speed of the motion of the surface would be typically 1100 cm/sec.
During 19789 David Robertson and I repeated the configurational experiments using five separate strain gauge specimens; Stephen North was the subject. The data bear out the original conclusions that the most usual configuration of a surface of action is that of a vertical plane passing through the body of the subject. They are summarized in Table 5.2. Since Stephen's action takes place at rather smaller distances than that of Nicholas, the experiments were on a scale between four and five times smaller. The use of five sensors made it possible to define the plane of experimentation in more detail.
For the analysis of the new data we defined a synchronism ratio s for an event as the ratio of the number of synchronous signals to the total number of sensors exposed. The mean value of s, denoted , was calculated for a session. We also tabulate a mean value weighted according to signal magnitudes. It is clear from Table 5.2 that both and decrease in the order RV > RH > EV. The surface of action is similar to that in the first series of experiments.
Stephen behaves differently from Nicholas in sessions, in that he cannot rid himself of the idea, which seems to be a correct one in his case, that metal-bending action usually extends from his hands or arms. Normally he points one hand, or even one finger, in the direction of the sensors. When these are mounted in a radial vertical configuration, it is natural that there are synchronisms on several sensors; the 'surface of action' appears to be an invisible vertical extension of Stephen's arm.
The occurrence of equidistant (E) sensor synchronisms in Stephen's sessions was therefore of particular interest. It became apparent to me during my observation of them that possibly both hands might be involved simultaneously. As will be seen from Table 5.2, the horizontal distances between the synchronous sensors were quite small; although Stephen was asked to produce action on the entire array, he was accustomed to point his left hand at the left-hand sensor; on occasion the right hand would also point, as though he felt that this was the natural way to produce a wider action. When both right and left hands were pointed, synchronisms would sometimes occur at the sensor at which they were being pointed. Thus it is possible that more than one surface can be produced by one subject.
Table 5.2
Session Configuration Number of sensors Horizontal or radial extent (cm) Vertical extent (cm) No. of signals mean s weighted mean s
EE RV 5 26 26 14 0.69 0.90
PP RV 4 21 13
GG RH 5 26 7 19 0.41 0.49
LL RH 5 30 8
FF EV 6 36 15 33 0.31 0.29
HH1 EV 4 18 15

A pair of remarkable sessions was held with Stephen using one sensor strapped to the forearm and one suspended in front of him. The forearm sensor was in the form of a circular disc with a rosette of three strain gauges at the centre, with this equipment the direction of the individual strain vectors can be determined (chapter 10). The disc was mechanically decoupled from the forearm by being mounted only on its screened leads, so that it was raised about l cm above the hairs on the forearm. If signals were obtained on this sensor, the experiment would show to what extent the strain vectors were aligned along the forearm. Also there was the question of whether synchronous signals would be observed on the forearm sensor and the suspended sensor.
The data from these sessions do not support the notion that the forearm sensor strain vectors show any tendency to be aligned along the forearm; the angular distribution appears to be fairly random.
However, the arrangement in time of the dynamic strain signals on the forearm and suspended sensors turns out to have tantalizing features; each set of signals on the forearm is followed after an interval by a signal on the suspended sensor. Of course this is only one of several possible interpretations, but it is nevertheless worthy of notice. What is surprising is the very long series of times between the corresponding signals. If this interpretation is correct, a very slow speed of the surface of action is indicated.
The notion that the surface of action passes through the subject (in opposite horizontal synchronous events) is somewhat imprecise. Are some part or parts of the body involved, and are they always the same parts? There is a common belief that the hands are involved. There is a case of a boy obtaining bends by placing the specimen outside his thigh, and success has also been achieved under the armpit. Two metal-benders are reported to have produced bends with their toes and two have used their foreheads. The neglected but classic experiments of Crawford(24) on table-lifting phenomena in the Goligher family led him to conclude that a 'cantilever arm', which was similar to our 'surface of action', came out of the lower half of the body, often the feet. Perhaps there are extensive differences between different subjects. More experimentation is needed; one series of experiments of my own was carried out not with strain gauges but with orthodox psychokinetic apparatus: a pointer suspended within a glass dome (chapter 20). One subject made detailed attempts to sense the parts of his hands and arms (e.g. the acupuncture points) which were most effective when placed under the suspension; but no special points were found. My experience with metal-benders suggests that what is important is the unconscious mind; the part of the body is dictated by psychological rather than by physical considerations.
A question often asked is: 'Does the right hand act more effectively than the left in a right-handed subject?' We have carried out a quantitative experiment with Julie Knowles, an ambidextrous subject with some right-handed bias. It is necessary that both hands should produce detectable action over the same period of time, in order that time-dependent variables (such as learning or decline effects) be eliminated. For example, strain gauges could be mounted on two metal specimens, one in the vicinity of each hand. An even simpler experiment was attempted with Julie, in which a long metal specimen (3 mm X 1.3 cm X 70 cm) was grasped with both hands, mutually separated by about 15 cm, under video-camera observation. It was found that an S-bend was produced, and at 5-minute intervals the magnitudes of the two parts of the S were measured by tracing onto paper. The data showed that the action of the right hand was 'stronger' than that of the left.
We can also consider the question of the maximum distance at which metal-bending effects are still possible. I have never conducted metal-bending experiments in a large hall or in the open air, so that the limits of action have always been set by the small room in which the subject and the specimens were situated. Nicholas Williams was able regularly to affect strain gauges at 5 m distance, and on one occasion very large signals were observed in synchronism, with a distance of 9 m between them; one strain gauge was on the ground floor of his house and another on the third floor, where Nicholas himself was situated.
On several occasions during an afternoon with Nicholas, we left the strain gauges and chart-recorders switched on in a locked empty house and went for a walk together. It was possible that signals would be recorded while we were out, since Nicholas was intending that they should. But nothing of significance was recorded.
At present I believe that metal-bending is limited in distance to the psychological 'immediate surroundings' or 'territory' of the subject; perhaps to the room in which he is situated. Further 'distance effects' will be discussed in chapter 8.
During the sessions with Nicholas Williams some experiments were carried out with partial screening of one of the sensors. The screens, of glass, brass and mild steel, are drawn in cross-section in Figure 5.2. The procedure was to wait until synchronous signals were obtained, and then quickly surround one of the strain gauges with a screen. It was found that the synchronism was destroyed, while the strength of the signals remained unimpaired. The data of Table 5.1 demonstrate this finding, but the statistics are poor. I have conducted further experiments with Stephen North, which confirm that signals are not impaired in strength by partial screening with transparent plastic. Total screening by a sealed glass sphere seriously inhibits metal-bending action, as we have seen in chapter 3. But the partial screening serves merely to delay the action by seconds.
We can conclude this chapter with the generalization that reliable synchronous strain gauge data tell us much about the metal-bending action; the action is repeatable in that different experiments are similar, each to each. It takes place at a 'surface of action', which may be regarded as a kind of invisible extension of the subject's arm.
Figure 5.2 Cross-sections of screens drawn to scale.

Back to books list.
Back to main index.
Back to Uri Geller's home page.