The purpose of this study was to develop and test a computer imaging method for assessing longitudinal processor variability. A technique called adaptive histogram equalization was used to test deviation in automatic processing when a test film is compared with a time-temperature processed standard at two different kVp's and film speeds. In a 14-day study we found that the density ranges for the standard film did not change, while the density range of the automatically processed films changed by a factor of two. These results suggest that if automatic processing is to be used for films taken as part of a longitudinal study-for example, subtraction-then processor variation can potentially lead to incorrect inference of bone gain or loss.