The annals of scientific fraud have now been enriched by another episode, this one at the illustrious National Institutes of Health, where five published research papers on the genetic origins of leukemia were recently retracted with the explanation that they contain faked data. The culprit in this case is said to be a graduate student accused by his laboratory chief of ''a stunning series of data misrepresentations and outright fabrications, extending over a period of at least two years.''

Science, of course, is not exempt from the impulse to chiseling that afflicts other sectors of society. And it may even be, as the science establishment claims, that fraud in research is both extremely rare and bound to be detected in the normal process of scientists building upon each others' achievements. Nonetheless, the newly announced five-paper fraud, in common with many of its predecessors in scientific delinquency, does suggest that publish or perish has finally swept away all other considerations for career advancement. As a consequence, the intended audience for published research papers is so obsessed with producing and publishing its own papers that scant attention is given to reading, let alone understanding, what the other guys are publishing.

Worldwide, ever-growing legions of researchers are turning out soaring volumes of papers even as they lament shortages of money for research. In clinical medicine, the worldwide volume of papers rose from 116,371 in 1981 to 130,957 in 1993, according to the National Science Foundation. Over those same years, the output in physics increased from 45,561 to 63,789. The folklore of science says that anything can be published someplace. Readers defy enumeration. But as the volume of papers goes up, researchers complain they can't keep up with the literature. No matter, they add, since very little of it is worthy of attention. In any case, one reason they can't keep up with the torrent is that they're too busy contributing to it.

The question of who's reading the stuff was raised last summer in an aside by the appeals panel in the so-called Baltimore case, involving the accuracy of a research paper published in 1986. The paper, the panel of two lawyers and a scientist stated, ''is rife with errors of all sorts.'' Several had been corrected in the course of the long dispute over the paper, but the panel noted ''additional errors evident on the face of the paper, some of which, despite all these years and layers of review, have never previously been pointed out or corrected.''

Publication credits are highly coveted as symbols of accomplishment necessary for advancement. The pressure to publish has created paper mills throughout the scientific establishment. It has also led to nasty squabbling about the allocation of credit between juniors in the laboratory, who often do the real work, and seniors, who control money and promotions. In cases where the later demand credit without having made an intellectual contribution, they are snidely known in the trade as ''honorary authors.''

Several years ago, the elite National Academy of Sciences declared that honorary authors ''dilute the credit due the people who actually did the work and make the proper attribution of credit more difficult.''

Science and scientists would profit from a severe cutback in volume of publication. But they're hooked, and no one knows a way out.

For the full story: Log In, Register for Free or Subscribe