A critical examination of the reference X-ray powder-diffraction patterns of the serpentine minerals in the Powder-Diffraction File (PDF) database has revealed an unsettling situation. Most of the patterns in, or previously in, the PDF database are inaccurate, misidentified, or of poor quality. The PDF database is not a dependable tool for identifying the serpentine minerals, and has not been since the mid-1960s. This has serious implications for studies on serpentine minerals that have depended on the PDF database, particularly those by nonmineralogists doing health and environmental studies of chrysotile asbestos. In the current PDF database, lizardite-1T, carlosturanite, some amesite, and possibly some antigorite (but with inappropriate polytype symbols) can be identified. Only one of the many multilayer lizardites can be identified. The current pattern for chrysotile-2Mc1 (clinochrysotile) is of reasonable quality, but not the best, however the earlier patterns still in the database are so problematic that any chrysotile-2Mc1 identification must be considered suspect. Chrysotile-2Oc1 (orthochrysotile), and any mixture of serpentines cannot be identified using the PDF database. Until the reference serpentine patterns are corrected the PDF database cannot be considered a reliable identification tool. High-quality powder-diffraction patterns of the serpentine minerals have been published and can be rapidly introduced into the PDF database.© 2000 International Centre for Diffraction Data.