Robert Wood's 1904 Visit to Nancy: The Removed Prism and the Collapse of N-Rays

In 2007 the American Physical Society's newsletter, APS News, ran a short historical piece on the N-ray affair: a radiation announced in 1903, reported in hundreds of papers, and abandoned by almost everyone within months of a single critical letter in 1904. The piece is a retrospective account by a professional society, not a primary document. Its value is that it sets out, compactly, three things a reader needs: what the rays were said to do, how they were detected, and what happened when an outside physicist quietly altered the apparatus in the dark.

What was claimed

According to the APS account, the claim began with René Prosper Blondlot, a physicist at the University of Nancy. The article stresses his standing: he belonged to the French Academy of Sciences and had a good scientific reputation. He had been studying the polarization of X-rays when he came to believe he was seeing a different radiation altogether. His first report appeared in the spring of 1903 in the Academy's proceedings, the Comptes Rendus. He named the rays after Nancy.

The article places this in a particular atmosphere. X-rays had been discovered in 1895, and many physicists expected further radiations to turn up. It adds that the announcement carried national feeling in France, because X-rays had been the discovery of a German, Wilhelm Conrad Roentgen.

Blondlot regarded the rays as a form of light. The APS piece says he reported that they could be polarized, reflected and refracted. The list of properties then grew, and the article's summary of it can be sorted into three kinds of claim. The oddest is one of storage: a brick was said to hold the rays. A second group concerns where the rays came from. The sun, gas burners and metals were all reported as sources, while wood was reported not to emit them. The third concerns what the rays could get through: wood and metals were said to be transparent to them, and water to block them. The article does not say who reported each property or how it was tested.

Two further claims are attributed to named researchers. Augustin Charpentier, a professor of medical physics at the University of Nancy, reported emission from rabbits and frogs and from the human brain, muscles and nerves, and predicted a medical-imaging use comparable to that of X-rays, showing the outline of internal organs. Jean Becquerel, the son of Henri Becquerel, claimed that the rays could be sent along a wire. The article presents these as reports and predictions; it describes no working application.

How the rays were detected

This is the heart of the matter, and the article is clear that the effects were described from the outset as barely detectable. It names two methods:

• In the earliest experiments, the indicator was a small electric spark, whose brightness was said to vary slightly when the rays fell on it.
• Later, screens with a phosphorescent coating were used; these were said to glow a little brighter under the rays.

In both cases, as the account presents them, the measurement was a human judgment of a faint change in brightness, made in a darkened room. The article does not describe any instrument that recorded the effect independently of an observer's eye.

How far the claim spread, and who could not see it

The APS piece reports that within months many scientists, mostly French but with a few others, said they had seen the rays, and that hundreds of papers followed, 26 of them by Blondlot himself. It does not give a total figure or a bibliography.

It also reports that many other scientists could not reproduce the results and saw nothing at all. The defenders' answer, as the article gives it, was that those observers lacked sufficiently sensitive eyes for an effect at the edge of visibility. The consequence is worth spelling out: on that argument, a failure to see the rays counts against the observer and not against the rays, so no negative report could settle the question. The article says the physics community was divided, and that the positive observers appear to have believed sincerely in what they reported. It does not allege fraud.

The visit and the prism

Robert Wood of Johns Hopkins University, who worked on optics and electromagnetism, was among those who had failed to detect the rays in his own laboratory. The article says he was sent to France in the summer of 1904 to watch Blondlot's experiments, with the aim of clearing the matter up; it does not say who sent him. It also notes his reputation for wide interests and for pranks.

Blondlot and his assistants ran several demonstrations. In the best known, an aluminum prism was said to spread N-rays into a spectrum, and Blondlot said he could find that spectrum because parts of a phosphorescent strip looked a little brighter than the rest. Wood saw no such spectrum.

Because the room was dark, Wood was able to take the prism out of the apparatus without Blondlot or his assistant noticing. He then asked for the spectrum observations to be repeated. The article reports that Blondlot went on describing the same pattern he had reported with the prism in place.

That is the decisive observation in the account. If the pattern depended on rays dispersed by the prism, removing the prism should have removed the pattern. It did not, which points to the observer and not the apparatus as the source of the reported brightenings. The article says that after several demonstrations of a similar kind Wood was fully convinced the phenomenon was imagined; it does not describe those other demonstrations.

The letter to Nature and what followed

Wood wrote to Nature on September 22, 1904, and his report appeared in the issue of September 29, 1904. As the APS article quotes it, Wood had watched experiments for upwards of three hours and could not report "a single observation which appeared to indicate the existence of the rays". He came away firmly convinced that the small number of experimenters with positive results had been "in some way deluded". The article notes that Wood did not name Blondlot, though readers would have known whose laboratory was meant.

By the article's account, belief in N-rays had nearly vanished within months and the question was treated as closed. Blondlot did not concede error and continued working on the rays for years.

What this account does and does not establish

The APS piece is a secondary summary written about a century after the events. It describes both Wood's letter and Blondlot's first Comptes Rendus report but quotes only the letter, and it gives no account from Blondlot's side of what happened during the visit. The prism episode therefore rests here on Wood's version as relayed by the article.

It also leaves the mechanism of the error unexplained. The article treats the affair as a warning about self-deception in observation, but it does not analyse why so many trained observers agreed with one another, nor does it describe any blinded test other than Wood's improvised one. Those are questions for the primary documents: Blondlot's own papers and Wood's letter in Nature.

Further reading

• Sheldon Stone's "Pathological Science" (2000): N-Rays, Langmuir's Six Symptoms and the Threshold of Seeing takes up the question of mechanism left open here: a physicist's review that reports Pringsheim's test of the faint-light detection method and sets N-rays against Irving Langmuir's six symptoms of pathological science.
• The N-Ray Affair in Overview: Scale, Failed Replications, Priority Claims and the Historical Literature gives a figure for the literature (about 300 articles), names Nature as the journal that asked Wood to go, describes a second substitution during the visit, and lists the historical studies by Nye, Klotz and Lagemann. Its section on finding the cited studies gives the identifiers for Wood's Nature report and shows which later work each claim about the visit depends on.