The transformation of digitalis from a folk medicine, foxglove, to a modern drug, digoxin, illustrates principles of modern pharmacology that have helped make drugs safer and more effective. Digitalis was improved because its preparation was standardized, first by bioassay and then by chemical methods; however, few of today's herbs are standardized by methods that can ensure a consistent product and, hence, consistent safety and efficacy profiles. Many herbs have been evaluated in randomized, controlled trials, and several-St. John's wort and ginkgo, for example-are apparently effective. Yet, many trials of herbs have limited value because of poor design, small samples, and, above all, use of products of uncertain composition and consistency. The uncertain composition of many herbal products raises questions about their safety, as does evidence indicating that herbs may have harmful interactions with prescription drugs. Such adverse effects of herbs are probably underreported. Meanwhile, systematic studies, such as those identifying adverse reactions to drugs, are hindered because herbal preparations are not standardized-one brand of St. John's wort, for example, will differ chemically from another-and, unlike for prescription drugs, there are no databases linking herb consumption to later medical problems. Since herbal medicines are regulated as dietary supplements, they are not subject to the premarketing regulatory clearance required for drugs. The burden of proof is on the U.S. Food and Drug Administration to show a dietary supplement is unsafe, unlike for drugs, which cannot be approved until the manufacturer has demonstrated safety and effectiveness.