In a darkened laboratory in Nancy in 1904, an American physicist secretly removed a prism from the apparatus his French hosts were using. They went on calmly reporting the spectrum the prism was meant to produce, as if nothing had changed. A celebrated discovery was about to collapse.1
An extra light in Nancy
The story began with an ordinary question. In February 1903, René Blondlot was studying X-rays at the University of Nancy. His apparatus included a tiny electric spark, whose brightness he watched as he changed the arrangement around it, believing the spark could reveal the orientation of the radiation reaching it. Turning the spark gap seemed to change the effect, and he took that as evidence that X-rays were polarised.2
Polarisation concerns the direction in which a wave oscillates. Blondlot was testing a specific physical relationship, not hunting for something strange in the dark. His February paper set out the wiring, the positions of the tube and the spark and the screens in between, instructions that another physicist ought to have been able to follow. The difficulty lay in judging what the tiny spark was actually doing.2
By 23 March 1903, his explanation had changed. A quartz prism seemed to bend the radiation towards the spark, and moving the prism seemed to change its brightness. Blondlot concluded that the effects came from a different kind of radiation from the X-rays he had set out to study, and announced a new species of light. The discovery began as a correction to his earlier interpretation. It never had a detector of its own.3
The radiation and the evidence for it were two different things. The proposed rays were invisible, so what the observer saw was an ordinary glowing object that supposedly responded to them, and a brighter spark counted as evidence only if the brightening could be reliably linked to the unseen radiation. Without that link, the experiment recorded a change in appearance and attached an explanation to it.3
An incandescent gas mantle soon joined the X-ray tube as a supposed source. In a paper dated 11 May, Blondlot described enclosing an Auer burner in a metal lantern, letting its radiation out through an aluminium window and concentrating it with a lens, with the little spark again serving as the detector. Water and some materials seemed to block the effect, while others seemed to let it through. The phenomenon was no longer confined to one complicated tube. It had become a programme of experiments with everyday objects.4
On 25 May, he named the radiation N-rays, in honour of Nancy. He also described new sources and detectors, including a small gas flame and calcium sulphide that had been made phosphorescent by light. Phosphorescence is the glow some materials keep giving off after the light that excited them has gone, and Blondlot reported that the new rays strengthened it. The material did not blaze into life. The observer had to notice a slight increase in a faint glow.5
The claim soon spread beyond Nancy. Writing shortly afterwards, the psychologist Henri Piéron described how a small problem in physics had become a controversy in which psychology had a stake too, tracing the stream of announcements that followed the first. Each new report offered another place to look for the same elusive effect. A disputed spark was growing into an entire experimental world.17
Bricks that kept the sun
By November 1903, that world could produce results after its source had gone. Blondlot reported that a phosphorescent detector stayed responsive after the burner was put out and carried away. The quartz lens, he explained, had stored the radiation and become a source in its own right, and other materials seemed to do the same. An effect that outlasted the apparatus became one more property of the rays.6
For Blondlot, a lingering effect was not an embarrassing failure to switch the rays off, but a discovery that an object had absorbed something and was releasing it. The idea could be tested, but it meant that removing the source no longer gave a clean comparison between rays present and rays absent, since anything near the experiment might now count as a source.6
A brick left in the sun supplied one of the most memorable examples. In a communication of 23 November 1903, Blondlot described bringing such a brick close to an observer looking at a dimly lit strip of paper. The paper appeared brighter. He went on to describe a clock face seen at a distance in poor light: with the radiation directed at the observer’s eyes, the vague patch seemed to resolve into a recognisable dial, hands and all. The rays were now acting on the person looking, not just on the thing being looked at.7
That raised an obvious problem. If water blocked N-rays, how could they reach the eye, which is full of water? Blondlot answered by distinguishing salt water from pure water, reporting that salt made water transparent to the radiation, and the explanation was part of a serious experimental argument. The question is whether his tests established it without relying on the same uncertain visual judgements.7
The field soon reached the human body. In a later French historical account, Jean-Michel Bloch described the work of Augustin Charpentier, a medical physicist in Nancy who attributed N-rays to animal tissues, contracting muscles and nerve centres. That was a long way from watching a spark beside a tube. It suggested that living bodies produced a detectable physical emission, and it opened new routes into the subject for other investigators.20
Blondlot was no unknown amateur seeking attention from a closed scientific establishment. Bloch’s account records his earlier academic work and the recognition he had won for research on electromagnetism, which explains why colleagues took his unexpected result seriously. It also changes the historical puzzle. The question is how experienced scientists came to trust an unreliable observation, not how they fell for an outsider’s trick.20
The equipment had a practical appeal too. A brick, a small glowing screen and a familiar clock made the claimed effect easy to imagine, and the demonstrations looked simple enough to repeat anywhere. But simple equipment does not guarantee simple interpretation. If the crucial change is still a marginal impression of brightness, adding more everyday objects only multiplies the opportunities for observation without making the signal easier to see.
Learning to see
Blondlot acknowledged that observers differed. In a November paper, he explained that some people saw the changes at once, while others at first found them at the very limit of perception, and that practice could make the effect clear. He also described ordinary reflected light, from a strip of white paper or a polished knitting needle, as responding to the rays. The spark was no longer needed, but the observer still had to recognise a slight change in brightness.8
There is nothing absurd about a trained observer getting better at a delicate task. The hard question is what shows that training has improved the detection of something real. If a learner is told what should appear and then learns to report it, agreement proves nothing. A useful test has to tell better observation apart from better anticipation, especially when teacher and pupil both know how the apparatus has been set up.
Another paper explained why people looking at the same screen might disagree. Blondlot claimed that the rays increased the light seen from straight ahead, reduced it when viewed at a sharp angle and made no noticeable difference from somewhere in between. An audience could not simply gather round and expect to see the same thing. Different viewing positions became a physical explanation for contradictory observations.12
Blondlot warned explicitly that an audience would see contradictory results, or nothing at all. A visitor could stand beside a successful observer and still fail to see the effect. The explanation preserved a physical reason for the disagreement, but it also limited how persuasive any demonstration could be. A clear test needed observers to take the specified position in turn, without knowing the state of the apparatus.12
The precision of Blondlot’s papers made them hard to dismiss. In January 1904, he described an aluminium prism dispersing the rays, with a narrow phosphorescent strip used to locate the separated components. He calculated refractive indices from their apparent positions and compared results from prisms and lenses. Refraction is the bending of radiation as it passes between materials; dispersion separates components that bend by different amounts. These were standard methods, applied to a signal nobody had confirmed.9
In February, Blondlot added N1-rays, a variety said to dim a weak source instead of brightening it. His paper describes searching the less-deflected part of the spectrum after another investigator’s observations suggested the idea, and finding the expected dimming, which disappeared when the rays were blocked. Brightening and dimming could now both belong to the programme, as long as each was assigned to the right kind of ray. Telling them apart depended on exactly the visual judgements already in dispute.11
Historians have placed that problem within the changing science of measuring light. Sean François Johnston describes a period when visual and physical methods were being compared and reworked, not one in which everyone agreed that the human eye was useless. An observer’s eye could be part of a serious instrument. The N-ray dispute turned on how to hold such observations to account, and whether another method would confirm the claimed change or simply inherit its weaknesses.22
A photograph of a flicker
Photography seemed to offer a way out. Blondlot’s February 1904 paper was careful to distinguish between photographing N-rays and photographing their supposed effect. The rays were not meant to act on the photographic plate directly. They brightened a small ordinary spark, which the plate recorded. If the image changed when the rays were let in, that would leave a permanent trace anyone could inspect, without learning how to look in the dark.10
His method alternated the exposure of two halves of a plate, with a screen moving alongside the plate holder to switch the spark between conditions with and without the rays, while a metronome timed the intervals. Alternating the conditions was meant to cancel out random fluctuations, answering an objection with a procedure, so the argument moved on to whether the procedure isolated the cause.10
In June, John Butler Burke at the Cavendish Laboratory in Cambridge challenged exactly that. He had tried photographic tests of his own and found no difference that could not be put down to an ordinary source of error. A metal screen placed close to the spark, he suggested, might itself change the electrical conditions; he proposed removing the source altogether, or keeping the screen further from the spark. A barrier meant to block one influence could introduce another, producing a difference for the wrong reason.14
Burke also separated an effect on a glowing object from an effect on an observer’s eye. These were different claims: a change that happened only in perception would not make the object’s light mark a photographic plate more strongly. Gathered under one name, the two explanations sounded mutually supportive, and Burke exposed the tension between them. If photographs were to settle the question, they had to show a real change in emitted light, with its ordinary electrical causes controlled.14
In America, R. R. Ramsey and W. P. Haseman reported their own photographic attempts. In some arrangements, they watched a phosphorescent screen fade and found no sign that the supposed rays slowed the fading. Apparent differences on other plates were tested further and traced to heat or to how brightly the source was glowing at the start. Their report also included spark photographs that showed nothing.13
A plate records whatever light reaches it while it is exposed. It cannot say why the light arrived, whether the source flickered or whether the comparison changed something else. A difference in a photograph can be real while the new force invoked to explain it is absent. Moving from an impression in an observer’s eye to a mark on glass does not make the question of evidence go away.13
The photographs survive. MIT Libraries’ discussion of Blondlot’s collection reproduces the paired images he offered as proof, and notes the objections later raised about how they were made. They let us see what looked persuasive at the time instead of simply being told that earlier researchers were fools. They are evidence of an attempt to demonstrate N-rays, not evidence that N-rays existed.19
An unwelcome visitor
The visitor who made the affair famous was Robert Williams Wood, a physicist at Johns Hopkins University in Baltimore. The American Physical Society’s history notes his work in optics and his reputation for experimental ingenuity and practical jokes. In 1904 he travelled to Nancy to see the demonstrations for himself, which helps explain his starring role in later retellings, where the dark room is both laboratory and stage.18
Wood told *Nature* that he had arrived sceptical but hoping to be convinced, and that after more than three hours of demonstrations nothing had persuaded him. His letter appeared on 29 September 1904, and it did not name Blondlot.118
He had asked an observer to say when his hand was blocking the rays reaching a spark. The reported changes in brightness bore no relation to his movements, and changes were still announced while his hand was perfectly still. The observer could not tell when the supposed cause was present.1
The strength of the test lay in keeping the information apart. One person knew what the hand was doing, and another judged the light. A convincing result required the two records to match, which takes far more than a shared impression that something brightened after someone announced an intervention. An experiment could convince its operators and still fail as soon as a visitor changed what they were told.
There were theoretical doubts too. In a later study, the philosopher János Laki compares objections concerning the reported wavelengths, the optical arrangement and whether closely spaced components could really be separated. He explains why declaring the results incompatible with existing physics could never settle the matter for supporters: a genuine discovery might require existing ideas to change. What it could not do without was an observation whose reliability did not depend on believing in the discovery first.23
The critics’ alternative explanations differed, with some focusing on perception while others looked for ordinary physical causes such as heat. Both could be right at once, if a screen genuinely changed in one arrangement while another demonstration produced nothing but an expected impression, but neither would establish the radiation Blondlot described. Finding that the reports had several mundane causes would complicate the story without rescuing the name under which they had been published.23
Once Wood’s hand test had failed, no new account of what the visitor ought to have seen could repair it. The question now was whether the apparatus could give a discriminating result, one that differed when the cause was present and when it was absent. An experiment that gives the same answer either way supports no claim about causes, however experienced the observer.
The prism in the dark
The aluminium prism offered a sharper version of the same test. It was supposed to bend and spread the rays so that an observer could locate the components of a spectrum with the glowing detector. The prism was not the source of the rays. It only changed their paths. Take it away, and there was no reason to find the same spread of components in the same places, even if radiation from the source still passed through the apparatus.9
During the demonstration, Wood secretly removed the prism, and the observers went on reporting the spectrum. The American Physical Society’s history preserves the episode, which has become the standard explanation for the collapse of N-rays, but its force lies in the unchanged observations, not in the cleverness of the trick. The experimenters were describing the effects of a component they did not know was missing, so their readings did not depend on the arrangement they thought they were measuring.18
Wood also reported three failed attempts to identify which way the prism was facing, which his hosts put down to fatigue. In the clock demonstration, he secretly swapped a steel file for a piece of wood, and the reported improvement in visibility continued. The prism was one of several failed comparisons.1
Fatigue can affect performance, and three trials are not a modern statistical study, but the point is narrower. A failure under those conditions cannot support a claim that the observer detected the rays under them, and if tiredness made the readings unreliable, a reliable reading still had to be shown. The excuse might explain a failed test, but it could not turn the failure into evidence for the rays.
Nor did Wood’s tricks reveal what anyone was thinking. Someone reporting an effect after the prism had gone might sincerely believe they could see it. The test examined the relationship between apparatus and observation, not people’s intentions, and that failure is why the episode counts as a scientific error. Calling every participant a conscious fraud would need different evidence entirely.
The decisive comparison is between an open test and a concealed one. In an open demonstration, the observer knows what the operator has changed and describes the expected effect. In a concealed test, that knowledge is withheld, and the observer has to detect the change independently. Success in the first followed by failure in the second makes expectation the obvious explanation. It does not mean nobody saw any light, or that every flicker was imaginary.
The scene is memorable because the missing object makes the problem visible. Months of apparatus, instructions and calculations had built up around the rays. A visitor removed one piece of the claimed physics while leaving the operators’ knowledge unchanged, and their readings followed what they still believed was on the bench.
After the missing prism
Wood’s report did serious damage, but not every believer gave up the day it appeared. An English translation of Blondlot’s collected papers came out in 1905, and its translator, Julien Garcin, presented them as a model of experimental discovery and a work of interest to students of scientific method. Whatever their physical merits, that introduction shows confidence persisting well after the encounter now remembered as the end.16
The collection also reinterpreted earlier results in the light of N-rays. Its preliminary notice explained that work first described as measuring the speed of X-rays had really concerned N-rays, and presented the resulting agreement with the speed of light as confirmation. A research programme can absorb earlier observations into a newer interpretation like this, until rejecting it means more than withdrawing one isolated claim.26
A review in *Nature* by John G. McKendrick in June 1905 was far less welcoming, contrasting the positive reports with failures elsewhere and questioning whether a faint visual impression could count as proof. The photographs were not exempt either, since differences in exposure or development could explain their contrasts, so turning the papers into a book had preserved the argument without settling it.15
French opinion was not united either. Pierre Potier’s later account for the Association française pour l’avancement des sciences describes an inquiry by *La Revue Scientifique* that exposed disagreement among French researchers and calls for more decisive experiments. In his reconstruction, support dwindled unevenly, without any single formal verdict. The old joke that only French eyes could see the rays hides a more useful question: who accepted which demonstrations, and on what grounds?21
Local and professional loyalty may have lent a respected scientist’s claims extra weight, but it cannot tell us what any particular supporter privately wanted to believe. Inferring a national campaign from where the work was done would repeat the original mistake of treating a plausible explanation as an observed fact. The firmer ground is the dispute at the time over method: whether confidence in an observer should outweigh repeated failures to confirm the effect independently.
Later research has also challenged the dramatic biographies attached to the affair. The physicist Robert T. Lagemann went back to Nancy, consulted local records and interviewed a former assistant who had known people involved. His account disputes the misleading picture of Blondlot’s later years found in some retellings. Memories collected long afterwards need their own caution, but nothing in them supports the neat invented ending in which a single exposure destroyed a man’s mind and life.24
None of this hints that the rays might one day be vindicated. The evidence for them failed. The open questions concern individual observations, the weight of particular influences and how a failed programme lost its supporters, which is to say how the episode happened and how it has been remembered. N-rays have no place among established physical phenomena.
The record that survived
What remains can still be examined: the original papers, their translation, the photographic claims, the contrary experiments and the letter describing the missing prism. Together they show the argument developing stage by stage, not just the polished tale of a debunker defeating a delusion. They also contain discrepancies and later revisions, and reading a collection in order of publication is not the same as assuming every passage was written on the date printed above it.
The dates themselves correct the familiar legend. The rays emerged through a series of communications in 1903, their properties multiplied, and independent criticism had begun before Wood’s letter appeared, while the English book and the hostile review both belong to 1905. There was no final confrontation after which everyone agreed it was over, and the record preserves a long, drawn-out disagreement alongside the experiment that gave later storytellers their central scene.5141516
Johnston’s history of light measurement helps explain why the affair mattered beyond Nancy. The eye was an established instrument for comparing brightness, but efforts to standardise light measurement were changing the relationship between human judgement and physical recording, and N-rays sat right on that boundary. Replacing a person with a photographic plate did not remove the choices about exposure and apparatus. What mattered was a procedure that could expose an error, whoever carried it out.22
Piéron’s early account gave the supposed rays a new home in psychology, a classification that followed from the failure to establish them as radiation. It is not a diagnosis of everyone involved, nor proof that one psychological mechanism operated in every trial. A historian can find evidence of expectation shaping a procedure without knowing what each observer consciously saw on the screen.17
Nancy keeps a quieter connection. According to the city’s tourism office, Parc Blondlot, near the railway station, stands on land the family left to the town, including the grounds of the former house. It is a legacy entirely separate from the experiments, and a reminder that the name survives as more than shorthand for an error.25
As of 7 November 2026, nobody has recovered Blondlot’s radiation, and its existence remains unsupported. Uncertainty about the cause of every reported flicker does not supply the missing confirmation. The old papers keep the measurements, the precautions and the confidence with which the discovery was described. The later experiments show why that confidence could not save it.131521
Linked case files
Sources26
- The n-Rays R. W. Wood · 1904-09-29 · Accessed 2026-11-07
- On the Polarization of X Rays R. Blondlot · 1903-02-02 · Accessed 2026-11-07
- On a New Species of Light R. Blondlot · 1903-03-23 · Accessed 2026-11-07
- On the Existence, in the Rays Emitted by an Auer Burner, of Radiations Which Traverse Metals, Wood, etc. R. Blondlot · 1903-05-11 · Accessed 2026-11-07
- On New Sources of Radiations and New Actions Produced by These Radiations R. Blondlot · 1903-05-25 · Accessed 2026-11-07
- On the Storing of N Rays by Certain Bodies R. Blondlot · 1903-11-09 · Accessed 2026-11-07
- On the Strengthening Action of a Beam of Light on the Eyes R. Blondlot · 1903-11-23 · Accessed 2026-11-07
- On New Actions Produced by N Rays; Generalization of Phenomena Already Observed R. Blondlot · 1903-11-02 · Accessed 2026-11-07
- On the Dispersion of N Rays and Their Wave-Length R. Blondlot · 1904-01-18 · Accessed 2026-11-07
- Registration by Photography of the Action Produced by N Rays on a Small Electric Spark R. Blondlot · 1904-02-22 · Accessed 2026-11-07
- On a New Species of N Rays R. Blondlot · 1904-02-29 · Accessed 2026-11-07
- On Peculiarities Presented by the Action of N Rays on a Dimly Lighted Surface R. Blondlot · 1904-02-02 · Accessed 2026-11-07
- An Investigation of N-Rays R. R. Ramsey and W. P. Haseman · 1904 · Accessed 2026-11-07
- The Blondlot n-Rays John Butler Burke · 1904-06-30 · Accessed 2026-11-07
- Review of Blondlot’s N Rays John G. McKendrick · 1905-06-29 · Accessed 2026-11-07
- Introduction by the Translator Julien Garcin · 1905 · Accessed 2026-11-07
- Grandeur et décadence des rayons N Henri Piéron · 1906 · Accessed 2026-11-07
- 1904: Robert Wood Debunks N-rays American Physical Society · 2007-08 · Accessed 2026-11-07
- Year 44 – 1904: Rayons N MIT Libraries · 2011-02-19 · Accessed 2026-11-07
- René Blondlot et les rayons N Jean-Michel Bloch · 1986 · Accessed 2026-11-07
- L’affaire des rayons N Pierre Potier / AFAS · 2025-04-28 · Accessed 2026-11-07
- A Notion or a Measure: The Quantification of Light to 1939 Sean François Johnston · 1994-11 · Accessed 2026-11-07
- The One-Dimensionality of Scientific Relativism János Laki · 2019 · Accessed 2026-11-07
- New Light on Old Rays: N Rays Robert T. Lagemann · 1977-03 · Accessed 2026-11-07
- Parc Blondlot Destination Nancy · undated · Accessed 2026-11-07
- Preliminary Notice to N Rays R. Blondlot · 1905 · Accessed 2026-11-07