When a scientific method is used by an expert to reach a conclusion offered in court, the Frye ruling in 1923 and particularly the Daubert ruling in 1993 requires that the method itself has been shown to be valid. When applied to fingerprint methods, valid means accurately distinguishing between pairs of prints made by one and by two donors. Courts have ruled uniformly in more than 40 Daubert hearings since 1999 that fingerprint evidence rests on a valid method, referred to as the Analysis-Comparison-Evaluation-Verification (ACE-V) method. In this article, we discuss the scientific evidence needed to document the validity of ACE-V. We describe examples of experiments that would provide this evidence, and review the available published research. We briefly describe the testimony presented by fingerprint examiners in these hearings, intended to show that ACE-V meets the Daubert criteria for validity. We analyze evidence for the validity of the standards underlying the conclusions made by fingerprint examiners. We conclude that the kinds of experiments that would establish the validity of ACE-V and the standards on which conclusions are based have not been performed. These experiments require a number of prerequisites, which also have yet to be met, so that the ACE-V method currently is both untested and untestable.