For years, the notion that researchers should share data freely with fellow scientists has been discussed widely. Some argue that this is especially true when the data are generated in federally funded projects. A recent provision in the reauthorization bill for the National Institutes of Health (NIH) would ensconce this principle in law. NIH grantees would be required, on demand, to furnish their data to other researchers. According to the proposed legislation, research data would have to be preserved and made avilable for 3 years after the completion of a project, and for 5 years following publication of the results in a scientific journal. One objective, according to a congressional aide, would be to make it easier for scientists with dissenting views to obtain and reanalyze data collected with public funds. It would also facilitate publication of alternative analyses. Personal records and patent applications would be exempted, but other data from clinical, behavioral, or epidemiological research focused on the evaluation or efficacy of a drug, medical device, or treatment of any sort would be covered immediately. Is such mandatory sharing of data ethical and appropriate scholarship? Does it raise potential for abuse? Should colleagues be required to provide such access to data, whether or not federal support was used in its collection?