Information technology is hot. The HMS Beagle itself rides upon a river of rapidly accessible technical material that every day flows further into a Mississippi delta of linked sites. But what about the information itself? What about all those measurements, charts, plots, trials, computations, assays, and explanations that account for the bulk of scientific traffic across the Internet? Is the message evolving along with the medium?

Not really. Whether read from a phosphorescent screen or from a paper page, most scientific data is still bundled into the traditional article format, including title, abstract, presentation, and references. And in this respect scientific communication has changed very little in recent times, except perhaps to become ever more technically narrow in scope, a development related to the historic growth in the complexity of science and the consequential need to specialize.

Narrowness, especially as it relates to the language used in scientific journals and to the demands made upon readers, was the subject of a recent meeting at the University of Chicago. Most of those present were physics journal editors, but I would say that much of the discussion would apply to the life sciences and indeed to all of science. I shall consider some of the principal themes of that meeting.

Is there a communications problem among scientists? Many complain that journal articles, at least those outside one's own research area, are unreadable, but is that necessarily a bad thing? Don't scientists read journals, consult colleagues, serve on committees, attend meetings, teach genetics or quantum mechanics to the next generation, and deliver talks at Tuesday-afternoon colloquia? Don't they choose their own level of involvement with other scientists? As evidence of success, are we not blessed with a growing inventory of labor-saving (even life-saving) devices, procedures, and smart materials? Isn't this enough?

Probably not. For one thing, the conditions of science keep changing. New scientific tools come into being, but so do new scientific problems. Traditional research areas - particle physics, organic chemistry, virology - will continue to be important. But clearly new problems demand new solutions. Multidisciplinary problems - acid rain, the physics of DNA, the imaging of viruses at the atomic level, neutralizing hazardous wastes - demand multidisciplinary cooperation. Like the search for petroleum in fields already depleted of easily obtained surface oil, progress in science will require new methods of extraction. And this will require new ways of talking.

Robert Watson, who has worked at NASA, the White House, and the World Bank, is an expert on ozone depletion and has just taken on the immense task of running the Intergovernmental Panel on Climate Change. Pondering the role of science and scientists in shaping global society in the next century, Watson pointed out (at a recent physics meeting entitled "What Do Physicists Owe Society?") that what we need is not generalists but experts, experts who are patient and alert enough to listen to experts in other fields. In other words, scientists will have to do more talking and more reading to keep up.

To produce important results in their chosen research field. To be esteemed by their professional colleagues. To secure a stable source of funding. Perhaps to be appreciated by students, nonexperts, and even the general public. Actually, what scientists and all busy people really want is more time. To write well, to speak well, to listen well takes time. Unfortunately, time is not plentiful but scarce, and in a busy schedule filled with research commitments, family duties, and teaching responsibilities, good communication is often seen as a luxury.

According to Watson, however, it is a positive necessity if we are going to solve the scientific problems to come. Much can be done to mitigate the science communications problem, beginning at the graduate and undergraduate level, by inculcating good writing skills. Requiring more expansive writing in lab reports, say, or the institution of history-of-science courses including extensive written assignments and critiqued oral presentations, would help.

But a more substantive arena for addressing the entrenched problem of bad language is the science journal. Few read any journal cover to cover. Many hardly read articles outside their own area of expertise. Some scarcely do more than glance at the table of contents. How can nuclear physicists who haven't the time to read about atomic physics be persuaded to read about the physics of the brain?

Making journals more reader-friendly is not a new issue, and has already been addressed in a number of ways. At some journals, for example, clarity is deemed to be not a minor but a major condition for acceptance of an article. Unfortunately, this precept is often difficult to carry out in practice. At other journals, the staff do a fair amount of rewriting, but this is a luxury few editors can afford. No, for any reform to have a chance of being effective, it must be concrete and practicable. My suggestion is to start with something small. Start with the abstract.

All too often the typical journal abstract is a dehydrated, inhumanly distilled chronicle of facts and numbers expressed in extremely technical terms. It's as if the author were speaking Latin to a congregation of thousands (the readers of the journal), but being understood by only 30 or 40 of those sitting in the front two rows. What is an abstract for? Some will argue that an abstract should be a condensed version of the whole article, and that therefore it must precisely deliver, in full gory detail, the prevalent facts and important numbers. Inevitably, this dooms the article to go largely unread.

I would argue, in contrast, that the 30 or 40 experts are going to read the article anyway, on the basis of the author's name and the title of the paper, and that the abstract should serve another purpose. It ought to be an invitation to the reader, to all the readers of the journal, concisely explaining the problem at hand and summarizing the way in which this paper achieves a certain end. As in newspaper journalism, the abstract, perhaps in the first sentence or two, will have to convince the busy consumer to continue reading. In such an abstract, obstructive details must tactfully be left out. The abstract should offer an incentive, a plea, an act of persuasion for (nonexpert) readers. Like a politician seeking votes, the abstract will have to shake as many hands as it can in a very brief time.

Especially now, when abstracts lead a secondary life as electronic archival material in their own right (a sort of shadow journal), they should be more discursive. Dare I say even playful? Abstracts should pique one's curiosity, not rebuff us with a formidable show of technical detail. Writing such a mini-essay will not come easily to many busy scientists. But as any English teacher can tell you, asking a student to write a brief outline of her term paper is one of the best ways of forcing her to think more clearly about what she means to say in the paper.

The entertainment value of scientific journals should not be overlooked. The secret of the formation of galaxies may well be related to the folding of DNA molecules; the cure for AIDS might somehow arise from the study of quark interactions. But how are we going to make such discoveries if scientists aren't reading each other's articles?