Decreasing the length of grant awards for biomedical research has been suggested as a means of ensuring greater accountability. An analysis of grant programs of the National Institutes of Health and, in particular, at the National Cancer Institute, revealed that the length recommended for grants is closely related to their perceived scientific merit. A principal conclusion is that selectively increasing the length of grants for only the most outstanding applications competing for renewal might improve research productivity while reducing research costs and easing the growing burden on the peer review process.