Observations placeholder

Originally published in the Journal of  the Society for Psychical Research Vol. 51, No. 792, October 1982, pp. 343-367. DOWSING: A REVIEW OF EXPERIMENTAL RESEARCH  by GEORGE P. HANSEN* 

PARAPSYCHOLOGICAL INVESTIGATIONS 

   Since early investigations by the SPR, various investigators have approached dowsing from a parapsychological standpoint, notably at the Parapsychology Laboratory at Duke University. As with the biophysical work, various levels of control were used, and results have not been consistent. 

   J. B. Rhine (1950) tested Henry Gross to determine if he could discern whether or not water was flowing through an underground pipeline. The control valve was some distance from the dowsing site and was out of sight of the dowser. For some of the trials a coin toss was used to determine whether the water was to be turned on or off; for other trials it was determined mentally by the experimenter. 

   In the first section of the experiment, significant missing was found (p < 0.001), and the consistency of the missing was also impressive. A noticeable decline effect was also noted between the first and second halves; however, it was not quite significant. Rhine felt that the results indicated the use of ESP and discussed this later (Rhine, 1952). Several weaknesses of this pilot study should be noted. Possibility of sensory leakage was not completely eliminated. Slight vibrations from water moving in the pipe might have been present, and verbal cues between experimenters might have been possible. Although several impressive statistically significant items were found, it seems they were derived from post-hoc analysis, and as discussed by Nicol (1955), this might be explained by possible nonrandomness of the experimenter’s mentally determined trials. Taking all the data presented of the tests with Henry Gross, there was no overall significant hitting or missing; Rhine did not comment on this. 

   Remi Cadoret, an M.D. on the staff of the Duke Parapsychology Laboratory, also investigated dowsing. He experimented to determine if the results and patterns of responses on one ESP task could be used reliably to predict results on another similar task (Cadoret, 1955). 

   Several pilot series were first run. A penny was placed under one of 25 tiles arranged in a grid (5 by 5). The subjects were taught to use a pendulum (a button on a string) as a dowsing tool and were asked to pick the row and the column containing the penny. Although the possibility of sensory cues was not completely eliminated, overall results were not significant. However a very significant decline effect was found between the first and second halves (p = 0.0014). 

   In another experiment a map was laid over the 5 by 5 grid. The map showed the squares corresponding to the tiles underneath and also corresponding to a grid established in the back yard of the experimenter. For some of the trials, a penny was placed under a tile beneath the map; for other trials a penny was placed in one of the back yard grid squares (the subjects did not know whether the penny was under the map or in the back yard). For this experiment each subject made 18 responses, six using the pendulum, six using roller bearing dowsing rods (apparently L-rods), and six with the subjects calling his findings aloud. From the results of the tests with the pennies below the titles, Cadoret noted the pattern of response in relation to the target (e.g. some dowsers might consistently miss the target by one row or column). He devised a mathematical procedure to predict the correct target square given the dowser’s response, and was able to predict the location of the pennv for those trials in which the penny was in the back yard significantly better than chance, p = 0.018. This experiment was the most tightly controlled, and sensory cues seem to have been completely ruled out. 

   Karlis Osis also carried out several dowsing experiments while working at the Duke Parapsychology Laboratory (Osis, 1960). One of his special subjects was a Mr. Gwaltney, a superintendent with a local gas company. For one of the tests,  ten small trenches were dug and covered with boards. Eighteen inch long pipes were randomly placed in the trenches. The subject then walked over the trenches while holding a pair of dowsing rods. In the first part of the experiment, the experimenter was present with the dowser, knew the location of the pipes, and recorded the subject’s responses, but for the rest of the time the location of the pipes was unknown to the experimenter who recorded the responses. In the experiment with the targets unknown to the experimenter the results were marginally significant (p = 0.03). For the entire experiment, the results were quite significant (p = 0.003). Normal sensory cues could possibly have played a part; each time the targets were placed, there may have been some slight tell-tale disturbance of the surrounding area. The type of material of the pipe is not mentioned; perhaps the results could be explained by sensitivity to magnetic anomalies rather than psi. 

   Osis also conducted tests in the laboratory with two selected subjects; one had had spontaneous psi experiences, the other had previous dowsing experience. Either money or photographs were randomly placed under one of 25 tiles arranged in a 5 by 5 grid. Subjects were asked to indicate the row and column in which the target object was placed by using a button on a thread as a pendulum. Based on direct hits, the results were marginally significant (p = 0.02). Details given in the report are rather sketchy, and it is difficult to trace whether sensory cueing could have been involved. Osis also tested Cadoret’s hypothesis that the patterns of hits and misses on one task could be used to predict hits for a similar task; his data did not support the hypothesis. 

   Map dowsing tests were conducted at the Parapsychology Laboratory with subjects as far away as Germany. No significant results were obtained. Osis did find one subject with whom there were indications of a consistent missing pattern which might have confirmed Cadoret’s hypothesis, but not enough data were collected to draw final conclusions. 

   Two short studies were reported by Pope (1950) in Parapsychology Bulletin. Miss Kirby, Lecturer in Biology at Harrogate Training College in England, conducted a series of tests in which a special subject attempted to locate a coin placed under one of several possible pieces of thick cardboard. Although the experiment was rather short (63 trials), the results were very significant (p < 10-6), but though precautions were taken, sensory cues might have been available. In tests conducted by the Physiology Department at Guy’s Hospital London, dowsers were asked to determine whether or not water was flowing in a concealed pipe beneath them and to locate the courses of underground drains. Details are sketchy, and sensory cues were apparently not eliminated; they found quite positive results. Unfortunately few details are available, and no number of trials or successes was given. 

   Moss and Sands (1970) report an experiment in winch an experienced dowser was pitted against a novice and a person using a ‘scientific method’ to predict winners of horse races. The dowser held a pencil over the racing form until a pull was felt to the name of a horse. The novice attempted a similar method. In the first experiment the dowser ‘won’ more money than the other two when imaginary bets were placed but no statistical comparison was made in the report to determine whether the results were significantly different from chance. Recently, Anselmo (1978) reported a successful dowsing test. Subjects were asked to locate coins underneath poster-boards. The overall results were considerably above chance (p < 0.001). It is not clear whether sensory cues were entirely eliminated, and the experimenter recorded both the dowser’s guess and the actual location which allowed possible biased recording errors. 

   Francis Hitching (1978), author of the book Dowsing The Psi Connection (reviewed by Cox, 1978; and Hvman, 1979), conducted a map dowsing experiment testing Bill Lewis, a retired electrical engineer in South Wales, to see if he could use dowsing to locate ancient megalithic sites (standing stones, burial chambers, etc.) in North America. Lewis was given maps of several areas and asked to locate such sites. He held a pencil in one hand and a pendulum in the other; and he moved the pencil over the map until the pendulum indicated a good site. Lewis then asked himself a number of questions which could be answered yes or no. From the information so derived, he gave a description of the site. These predictions and locations were shown to John Stiles, chairman of the ESP committee of the SPR; he formulated a series of similar predictions for locations near the sites of those of Lewis. Stiles made his predictions by guesses based on those of Lewis. Hitching then visited most of the sites and compared the two sets of descriptions with the sites. He found much greater correspondence with the predictions of Lewis than those of Stiles. The major weakness of this procedure is the difficulty of making an unbiased evaluation. The descriptions given by Hitching did indicate that unusual results may have been obtained (his claims were far stronger). More rigorous controls could produce a more convincing demonstration. 

   A number of unsuccessful experiments have been reported. Mr. P. A. Ongley, a New Zealand research chemist, tested the claims of 75 dowsers. These ranged from medical diagnosis to tracking people, etc. He concluded that all were unwarranted. Although many tests were conducted, and a large amount of numerical data was presented in his article (Ongley, 1948); few details were given. Many of the claims appear to have been tested with only one or two trials. He seemed to make the tests fair to the dowsers, but the tone of his report suggests that he probably had a rather strong, preconceived opinion against dowsing. 

   Another series of unsuccessful experiments was reported in Nature by R. A. Foulkes (1971). Experiments were organized by the British Army and Ministry of Defence to determine if buried mines could be located by either map or field dowsing; dowsing for water was also tested. For the map dowsing tests, 20 inert mines were buried along several military roads. Seven dowsers were given maps of the roads and asked to locate the mines. Only chance results were obtained. For the field dowsing, a 20 by 20 grid was established with each square being 20 feet (6.1 metres) on a side. Five different types of objects (80 of each type) were buried randomly. Tests were conducted with 22 dowsers to determine whether they could identify the objects. Again only chance results were obtained. An experienced dowser was asked to determine whether water was flowing in a plastic pipe. The water was randomly turned on or off for 50 trials. No significant results were obtained. From the report, the tests seem to have been well organized and well conducted. 

   Another unsuccessful experiment was conducted by several members of The American Society for Psychical Research including Laura Dale, Gardner Murphy, and Montague Ullman (Dale, Greene, Miles, Murphy, Trefethen, and Ullman, 1951). Twenty-seven dowsers were taken separately to a small field near Liberty, Maine and asked to locate the best spot for a well, estimate the depth, and the flow rate. Pipes were later driven, water level measured, and the wells were pumped to determine the capacity. A water engineer and a geologist were asked to estimate depth and flow rate at several locations (the engineer and geologist knew of a nearby well, the dowsers did not). The soil was relatively soft, and the water table was nearly level and close to the surface. The geologist’s and engineer’s predictions were quite good; the dowsers’ predictions were quite far from the mark. Because the water table was nearly level over the site, perhaps there were no distinct subsurface anomalies. 

   Barrington and Stiles (1973) conducted an investigation of a commercial divining instrument called the ‘Revealer’. The Revealer was to be used for locating underground services and was basically a pair of L-rods. A number of public utilities, engineering companies, and local authorities had purchased Revealers, and Barrington and Stiles sent them a questionnaire regarding their use and satisfaction with the instrument. Most assessments were favourable. Five representatives of organizations giving favourable replies were selected for field tests. Several different testing procedures were used, but usually objects were buried in a sand pit, marked in a fashion to indicate a grid, and dowsers were asked to select the squares in which target objects (two-foot long sections of pipe of various materials) were buried. Subjects were usually given only 5 to 10 trials. Only one of the five persons tested showed promise. He was subsequently tested again (under admittedly poor conditions) but did not repeat his performance. 

   James Randi (1979), professional magician and member of the Committee for the Scientific Investigation of Claims of the Paranormal, conducted a test with four dowsers in Italy. Procedures were spelled out in detail prior to the test and agreed upon by the dowsers. The dowsers were asked to locate three buried pipes with running water and to place pegs over the route of the pipes. As stated by Chamberlin (1980), the test had several deficiencies. No meaningful statistical evaluation was possible. Even if the dowsers had been quite close, they were unlikely to fulfil the requirements for a successful test (they were required to place the pegs in a strip eight inches wide). None of them was able to claim Randi’s $10,000 reward. The test contributed little knowledge to the scientific community. 

   Bryant (1931), Carpenter (1877), Christopher (1970), MacFayden (1946), Parsons (1959), and West (1948) report unsuccessful tests. Unfortunately evaluations cannot be made since few details were given or few trials were conducted. J. W. Gregory (1928) and Sollas (1884) reviewed a number of tests conducted in the late nineteenth and early twentieth century; most produced negative results. 

   Overall, dowsers have performed reasonably well on dowsing tests purporting to require psi ability. Unfortunately reports are rather sketchy and some seem to indicate possible sensory cues. 

   Of the tests reviewed, only three stand out as well conducted—the map dowsing tests of Cadoret, Osis, and Foulkes. Of these, only one produced significant results, and that barely significant. It is unknown how many other  well conducted studies have been unsuccessful. Overall, the parapsychological investigations into dowsing remain inconclusive.

   If any additional work is done, it would be advisable to use map dowsing so that sensory cues could be eliminated (experimenters should be aware that some maps, topographical ones for instance, might give relevant sensory information, depending upon the dowsing task). Experimenters may get better results if definite steps are taken to create a positive environment and a realistic test situation as suggested by W. H. Jack.