Skip to Main Content
Skip Nav Destination

What it was like to be peer reviewed in the 1860s

9 May 2017

Personal letters, not formal referee reports, communicated suggestions for improving an 1861 paper on the physics of heat.

Tyndall experiment
John Tyndall used this experimental apparatus to compare how heat gets transferred via conduction and radiation. Credit: J. Tyndall/Phil. Trans. Royal Society of London 1861.

In January 1861 John Tyndall, a physicist at London’s Royal Institution, submitted a paper to the Philosophical Transactions of the Royal Society of London. The paper bore the title “On the absorption and radiation of heat by gases and vapours, and on the physical connexion of radiation, absorption, and conduction.” After testing the heat-retaining properties of several gases, Tyndall had concluded that some were capable of trapping heat, and thus he became one of the first physicists to recognize and describe that basis for the greenhouse effect. A month after its submission, the paper was read aloud at a meeting of the society, and several months after that, a revised version of the paper was in print.

That path from submission to revision and publication will sound familiar to modern scientists. However, Tyndall’s experience with the Philosophical Transactions—in particular, with its refereeing system—was quite different from what authors experience today. Tracing “On the absorption and radiation of heat” through the Royal Society’s editorial process highlights how one of the world’s most established refereeing systems worked in the 1860s. Rather than relying on anonymous referee reports to improve their papers, authors engaged in extensive personal exchanges with their reviewers. Such a collegial approach gradually lost favor but recently has undergone something of a resurgence.

George Gabriel Stokes and the Philosophical Transactions

Refereeing was not typical for scientific journals in 1861. Many commercial scientific publications, including Nature and Philosophical Magazine, accepted articles with little or no outside refereeing, and most journals on the European continent relied on the judgment of highly qualified editors to determine what would be accepted or rejected. British scientific societies often employed more involved review processes that gathered opinions from other experts in the paper’s field. Publication in the Philosophical Transactions conferred, if not the Royal Society’s endorsement, then at least its general approval, and the fellows of the Royal Society were eager to ensure that Transactions papers reflected well on their organization.

Papers submitted to the Transactions were routed to one of the society’s two secretaries, the de facto editors of the publication. The secretaries were eminent researchers who arranged refereeing and provided their own expert judgment on papers submitted for publication. One secretary handled the biological sciences, the other the physical sciences.

George Gabriel Stokes
George Gabriel Stokes. Credit: Smithsonian Institution

In 1861 the physical sciences secretary was George Gabriel Stokes, one of the most respected physicists in Victorian Britain. He held the Lucasian Chair of Mathematics at the University of Cambridge, the storied post once occupied by Isaac Newton, and made meaningful contributions to research on fluid dynamics and light. He was also a significant figure in the history of refereeing. When Stokes began his tenure as secretary in 1854, the Royal Society was still reeling from an intense public controversy about favoritism at the Transactions. To address such concerns, Stokes took on the task of standardizing refereeing for Transactions papers.

After receiving “On the absorption and radiation of heat,” Stokes arranged for two experts to review the paper. One was his closest friend, William Thomson (later Lord Kelvin), arguably the most important physicist in Britain. The other referee was Stokes himself. This was not unusual; Stokes had wide-ranging knowledge of physics research and often refereed papers in the physical sciences.

After a few weeks, Thomson returned a referee report to the Committee on Papers, the Royal Society group that made final decisions on which papers would be accepted. That report has unfortunately not survived, though Thomson’s correspondence with Stokes suggests that he recommended publication.

Stokes, on the other hand, did not file a formal report with the society. In a letter to Thomson, he mentioned that he “did not write a report on Tyndall’s paper but merely recommended it.” Although he told the committee that he disagreed with Tyndall’s conclusions about the relationship between conduction and radiation, he said that “if the author wished it retained he had done such good work that he had, I thought, a right to keep it.”

Stokes could make a verbal report to the committee in part because the referee reports, at that time, were not shared with authors—they were meant to inform the committee’s decision, not to help authors with revisions. However, Tyndall did receive feedback on the paper through another channel: personal correspondence from Stokes.

Stokes’s suggestions, Tyndall’s revisions

Stokes and Tyndall had known each other since the early 1850s. The two Irishmen likely met for the first time at a gathering of the British Association for the Advancement of Science, and their shared interest in physics led to a lifelong professional relationship.

John Tyndall
John Tyndall. Credit: Wellcome Library, London (CC BY 4.0)

In a letter dated 7 May 1861, Stokes informed Tyndall that “your paper was ordered to be printed last Thursday.” He told the author that he had been one of the referees, and he suggested five places where he thought Tyndall should clarify or reconsider his arguments. Some of the suggestions were minor. For example, Stokes rightly pointed out that heat could be lost to the rock salt in Tyndall’s experimental setup.

Stokes’s most significant comment was about Tyndall’s discussion of the relationship between conduction and radiation, the issue Stokes had mentioned to Thomson. Tyndall suggested that good conductors were, as a rule, bad radiators, and vice versa. Stokes told Tyndall that he did not believe it was necessarily a “physical law” that good conductors had to be bad radiators:

“You suppose that in the case of a good radiator a molecule does not lose much motion by communication to the adjacent molecules because it has already lost it by communication to the ether. . . . But it seems to me that the loss by communication to the ether and the loss by communication to other molecules must take place sensibly independently of each other.”

Tyndall’s published paper reflected changes in response to all five of Stokes’s suggestions. Most noticeably, Tyndall softened his conclusions about conduction and radiation and cast his work as an invitation to further experiments rather than a solution to the puzzle:

“Why should good conductors be, in general, bad radiators, and bad conductors good radiators? These, and other questions, referring to facts more or less established, have still to receive their complete answers.”

In making the changes, Tyndall does not seem to have been merely bowing to pressure from a powerful editor. In fact, he and Stokes exchanged several further letters about the paper. Tyndall sought and received clarification on several of Stokes’s criticisms. He clearly valued Stokes’s judgment and took his ideas seriously. In one letter, Tyndall told Stokes that he did not feel coerced to change his paper. Rather, “any point on which you have thought, and regarding which you have arrived at an opinion opposite to mine, demands from me very careful consideration before committing myself to print upon the subject.”

1861 versus 2017

Such open, collegial discussion will likely seem strange to modern physicists, who are accustomed to biting criticism from anonymous peer reviewers and to having limited opportunity to respond. There is much to admire, and maybe even envy, in Tyndall’s and Stokes’s exchanges about “On the absorption and radiation of heat.” On the other hand, although Stokes corresponded with many aspiring Transactions authors, including those he did not know well, we could easily see such a personal system favoring the editor’s closest colleagues—and putting less well-connected researchers at a disadvantage when trying to get their work published.

As the 19th century went on, refereeing at places like the Royal Society tended to become more formal and more systematized, with less room for the kind of exchange we see between Tyndall and Stokes. By the 1890s, members of the scientific community began envisioning a referee not as someone who merely made an expert judgment about the quality of a paper, but as someone who had to protect the trustworthiness of the entire scientific literature by detecting bad work. A more adversarial relationship between authors and referees developed, which meant that private discussions between them largely fell by the wayside.

Now, more than a century later, some new proposals for peer-review protocols have the potential to restore collegial discussion. Some observers have suggested that allowing readers to post comments about scientific preprints online would create a more open exchange between authors and referees—although critics point out that online comment systems are generally not known as bastions of collegiality. At the biology journal eLife, editors and referees consult and offer a single consensus report about a paper, rather than ask the authors to sift through two or three different sets of criticism (see Physics Today, December 2016, page 10). Those recent developments demonstrate that although peer review is a core component of scientific publishing, its form can change and has changed to adapt to the ever-evolving needs of the scientific community.

For a complete history of peer review in the sciences, see “In referees we trust?” (Physics Today, February 2017, page 44).

Close Modal

or Create an Account

Close Modal
Close Modal