Three physicochemical methods (HPLC, NMR spectroscopy, and HPLC of a derivative) have been used to measure parthenolide in authenticated Tanacetum parthenium (feverfew) and in several commercial purported feverfew products. A bioassay based on inhibition of the secretory activity of blood platelets by extracts of feverfew in comparison with parthenolide was also used. Similar results were obtained for all three physicochemical assays and also for the bioassay. Thus different methodologies yield consistent values for parthenolide content of feverfew preparations. Parthenolide appears to be mainly responsible for the antisecretory effects of extracts of feverfew. Authenticated Tanacetum parthenium grown in the UK contained a high level of parthenolide in leaves, flowering tops and seeds but a low level in stalks and roots. The level of parthenolide in powdered leaf material fell during storage. The purported feverfew products varied widely in their parthenolide content and in some products parthenolide was not detected. Possible reasons for the variation in parthenolide content are discussed. Since therapeutic efficacy has only been demonstrated for preparations of feverfew that contain parthenolide, it is suggested that manufacturers of feverfew products should use measurements of parthenolide as a means of standardization and quality control.