Transversal microradiography is the most widely accepted method used to study changes in mineral content profiles. In spite of its widespread use, relatively little information is available on its validity and reproducibility. Following the recommendation of the Consensus Conference on Intraoral Model Systems, this study was designed to explore reproducibility of lesion analysis within a laboratory and comparability of analysis among various laboratories. Incipient enamel lesions were produced by four research groups using both a common ('standard') and a local ('preferred') protocol. Sections were produced by each group and allocated to 'mixed' bags of specimens, which were analysed by the groups. With the chosen scheme some sections were analysed six times by the same group (as an internal reference standard) while others were analysed by all four groups. The data for the mineral content profiles were expressed as the integrated mineral loss (IML) value and lesion depth. The results showed the lesions produced with the standard protocol to be in the range 2,000-3,000 vol% mineral x microns for IML. The IML of the lesions produced with the preferred protocol varied between 1,800 and 6,300 vol% mineral x microns. Variation in IML values could be attributed to the biological variation between lesions, but also to time (of microradiograph production) and measurement effects, calibration of the magnification of the specimens, and the parameters used in the algorithm to calculate IML. Some of these parameters also affected the lesion depth. It is advised to standardise (or at least report) the method of calculation of IML, and to include a reference lesion between analyses in a longitudinal study as an internal standard. With the data produced, it was calculated that the number of lesions required to differentiate between preventive treatments varied substantially among laboratories. The recommendations given will improve the power of the screening methods for caries-preventive agents for which microradiography is an essential analytical method.