Nature insists on the retraction of papers it has published that are subsequently shown to be false. But sometimes that is easier said than done, especially when co-authors of the original papers disagree over whether there is any fault at all.

Such occasions are, thankfully, rare. This issue contains an example that is even more unusual. At the heart of a billion-dollar lawsuit that has just entered the phase of jury deliberation (see page 289) are the contradictory statements of scientists as to whether a plasmid discovered at and patented by the University of California at San Francisco formed the crucial basis of a critical experiment in the early commercial success of the company Genentech, Inc.

The conflict can be seen starkly in separate letters addressed to this journal (see pages 297-). Dennis Henner and current and former colleagues from Genentech deny misrepresentation in a subsequent Nature paper describing these experiments, and using stolen materials. In sharp contrast, their co-researcher Peter Seeburg confesses to what he describes as “a technical inaccuracy”.

Nature is in no position to arbitrate. It is now up to a San Francisco jury to decide who has made the most convincing case, and even that judgement is unlikely to be final. It is some consolation that neither the paper itself — nor the value of the human growth hormone to which it has led — are in dispute. But the picture of commercial pressures lurking in the background does little credit to the public reputation of the biotechnology industry — and makes the adoption of guidelines for sharing research tools between federally sponsored researchers and industry (see page 291) even more urgent.