Progress in science is congruent to the progress of the scientific community. Such a community can exist only through communication between its members. Perhaps one might argue that formal communication began to supplant informal processes with the birth of the first scientific journals, Philosophical Transactions of the Royal Society and Journal de Sçavans, in 1665. But the primacy of journals was slow to arrive. The young Willard Gibbs, newly appointed Professor at Yale, published his first remarkable scientific papers on thermodynamics in the Transactions of the Connecticut Academy, beginning in 1873. While that journal was not unknown in Europe, Gibbs insured reception of his work by augmenting the formal publication in a journal by the informal processes of sending copies of his early papers to Maxwell and other significant European scientists.
In this 1970s photo, Robert Adair is shown silver soldering apparatus onto a magnet for the 30-GeV Alternating Gradient Synchrotron at Brookhaven National Laboratory. Photo by Douglas Humphrey, Brookhaven.
From the problems of a paucity of formal communications, in this 21st century we have moved to problems of a superfluity of such communications. We have now a flood of communications and we risk the isolation of subdisciplines through the difficulties in discriminate selection from that flood.
Much of that move to selection was generated in the 19th century with the advent of journals devoted to one sector of science, such as the Physical Review, founded by Edward Nichols and Ernest Merritt at Cornell, which began publishing in 1893. The establishment of Physical Review Letters, by that remarkable prescient man, Sam Goudsmit, in 1958, marks a beginning of the further discrimination in the selection of communications which poses problems that still trouble us.
In the six bimonthly issues that made up that first volume of the Physical Review, covering the year from July 1, 1893 through the next June, there were 10 papers and 16 Minor Contributions, the precursor to Physical Review Letters, printed in 1 point smaller type.
In 1912, after 35 volumes, the Physical Review was taken over by the American Physical Society beginning a Second Series with Volume 1, January 1, 1913. Still a bimonthly, there were 34 articles and 1 Minor Contribution in that first volume. Through all of those early years, serious American physicists took a serious look at all papers even as they read all physics research papers in the major European journals.
Before 1925, the Minor Contribution section contained brief scientific notes, meeting reports, and book reviews. Then, in 1925, the journal began a differentiation with the scientific notes listed as Rapid Communications, and after 1929 the scientific notes were labeled Letters to the Editor, or just Letters.
In the following decades, the Letters were generally truly minor contributions, that is short notes on matters worth disseminating but not important enough to warrant a major publication. The basic information transfer in physics took place through the publication of articles—reasonably complete descriptions of significant research—in important journals. The Physical Review published about 200 such papers a year. Together with the major European journals, perhaps 500 papers a year presented physics of some reasonable note.
As a beginning graduate student commencing research in nuclear physics at the University of Wisconsin in 1947, my professor, Heinz Barschall, counseled me to read every month all articles addressing nuclear physics in the Physical Review as well as those in Nuovo Cimento, Helvetica Physica Acta, and the Proceedings of the Royal Society and to read the abstracts of all other papers in the Physical Review. And I did. I was told most of my elders had, in the pre-WWII decades, read all papers in the Physical Review but, with the expansion of physics, that was now becoming too time-consuming.
I was also told to join the American Physical Society and to subscribe to the Physical Review (at a reduced rate for graduate students). Impressed by the rows of bound volumes of the Physical Review in faculty offices, I extracted enough money from my veteran’s GI-Bill payments and modest teaching stipend to have my first years’ set (Volumes 73 and 74) bound, imagining that in the fullness of time, perhaps ten years in the future, I would have a set of twenty volumes of the Physical Review to display on my book shelves.
The Physical Review published 167 articles and 184 Letters in the 12 issues and 1477 pages of Volume 73. On looking through the authors of Volume 73, I am surprised to note that in the course of the next decade I got to know so many personally. Perhaps half of the papers had an author whom I later was to know personally on some level. The society of research physicists was within the human compass even as the reports of their researches were still universally accessible.
However, if the whole information flow was manageable—barely—in 1948, with 3366 published pages for the year, ten years later a close reading of the 7589 Physical Review pages was beyond practical reach.
After WWII the character of the Letters began slowly to change from their genesis as “minor contributions.” Some realized that discovery priority was established as firmly with a 600-word letter as a 6000-word paper. And the shorter Letters were handled more simply editorially—page proofs were not sent to authors and there was no review beyond that of the editor—and were thus published in a few weeks, much more quickly than articles. Moreover, with the length restriction of Letters, it was not possible, and hence not necessary, to describe all of the experimental details—trade secrets—to one’s competitors. Hence, as early as 1950, the Letters section of the Physical Review described many transparently important results.
Recognizing that important results were moving into Letters publication and that physicists now needed some accessible condensate, Physical Review Editor Sam Goudsmit, decided to publish the Letters section of the Physical Review separately. Thus Physical Review Letters—the first Letters Journal—was launched in July 1958, with Sam as Editor and the young theoretical physicist, George Trigg, as Assistant Editor. Trigg, later Editor, was a mainstay of the Journal over the next several decades. In a continuance of the former Letter procedures, the editors continued to limit the length of such submission to the new journal to 600 words and continued the procedures that led to publication times of less than a month; that is, the only review was that of the editor and authors were not sent proofs. In an editorial that opened the new journal, Sam wrote, “… only papers that really deserve rapid publication should be submitted.”
The twice-monthly Physical Review Letters printed 413 Letters in 1089 pages during the first full year. Sam wrote in an editorial that opened the first issue of that year, “We shall do our best to keep Physical Review Letters thin enough so that all of our readers can learn about all the significant developments in physics without having to wade through bulky volumes.”
With the promise of quick publication of summary results that would be very widely read, the pattern was soon formed of physicists reporting important results in a Letter, and publishing a detailed account later as an article—though often that completed accounting never came to pass.
With its fast publication of relatively brief accounts of important results, Physical Review Letters soon became a favorite journal for physicists to read—and thus to report the results of their research. As Editor of Physical Review Letters (1978-1983), I engaged in some correspondence with John Maddox, then Editor of Nature. In a letter to Maddox I described Physical Review Letters as one of the most important journals in physics. Maddox replied that it was “the single most important journal in all of science.”
Such success generated parallel results and Physical Review Letters’ European cousin, Physics Letters, began publication in 1962.
Part of the extraordinary success of Sam Goudsmit’s launch of a selective Letters journal came about from the ascendancy of English as the universal scientific language.
With the center of physics in Europe before WWII, Ph.D. students in physics were required to demonstrate a “reading knowledge” of appropriate foreign languages, usually French and German. In 1948 at Wisconsin I passed—somehow—reading proficiency examinations in French and German presided over by faculty in the language departments. Today, neither Wisconsin or my present school, Yale, require any language competence other than English for physics graduate students. With the increasing position of English as the universal scientific language, such foreign language requirements have faded away, further enhancing the position of Physical Review Letters. I find it almost poignant to read Zeitschrift für Physik today which is now wholly in English.
However, by 1980, the success of Physical Review Letters was such that the original model had to be significantly modified. The length limitations were extended considerably and submissions were submitted to outside reviewers who were asked to judge the submissions with attention to Sam’s criterion that the journal should only publish papers that really deserve rapid publication. Although Sam’s promise to keep the journal “thin enough” had eroded, his aim to publish “only papers that really deserve rapid communication” was kept as firmly as was practical.
However, even as the community of physicists, and their communications, was expanding, those communications were beginning to change in character. My correspondence with Maddox followed from some concerns he expressed that the Physical Review Letters acceptance criteria, published in 1982 and referred to in an editorial signed by Editors George Basbas, George Trigg, Gene Wells, and myself, did not require as a condition of acceptance that the contents of papers were not publicly available before publication in our journal. This position was partly driven by, and partly in anticipation of, the informal dissemination of important results that was becoming ubiquitous.
These first defections from the priority of journal publications were probably generated by the development of the Xerox machine. Although mimeographed preprints had been circulated sparingly since the early years of the 20th century, considerable effort was required to cut the necessary mimeograph stencils, and it was difficult to reproduce figures and some formulas. With the Xerox, ubiquitous after 1965, the production of high-quality copies of a paper was trivial and the dissemination of preprints expanded greatly. We Physical Review Letters editors understood that we neither could, nor should, attempt to interfere with this kind of prepublication information transfer.
This sans-journal information transfer, of emerging importance in 1980, exploded in the 21st Century along with the general explosion in electronic information transfer. The arXiv system, founded by Paul Ginsparg in 1991 as an electronic archive for preprints (later called e-prints) in theoretical physics, was especially important. Although a loose quality control is effectively imposed, arXiv papers are not reviewed. About this time, the American Physical Society began to put the carefully reviewed contents of all of its journals, from the beginning in 1893, into an electronic data base, PROLA, that was completed in 2001.
Now, some 50 years after Sam Goudsmit founded the journal, Physical Review Letters publishes about 3500 Letters a year. These published Letters are carefully reviewed and selected from a much larger set of submissions. Of course, that larger set stems from a physics community that is much larger than that which existed a half-century before. Thus, Physical Review Letters now publishes more articles and more pages than the whole Physical Review published in 1958. However, the selection still follows Sam’s 1958 criterion of “… only papers that deserve rapid publication ….” And while most physicists now scan Physical Review Letters online rather than in “…bulky volumes …,” arguably, John Maddox’s encomium is still valid; Physical Review Letters is still “the single most important journal in all of science.”
