A study was conducted to provide a replication of the gaze duration algorithm proposed by M. A. Just and P. A. Carpenter using a different kind of passage, to compare the three gaze duration algorithms that have been proposed by other researchers, and to measure processing time in reading. Fifty-one college students read a passage while their eye movements were monitored. Five different measures of processing time in reading were each fit, using hierarchical multiple regression, to a model similar to that of Just and Carpenter. The processing time measures--Just and Carpenter's gaze durations, two modified gaze duration measures, number of fixations, and average fixation duration--were shown to be influenced by different independent variables. Also, some evidence was obtained that called into question one assumption of the gaze duration measure--that when increased processing time is needed, a trade-off occurs between fixation duration and the number of fixations on a word. The findings suggest that gaze duration measures should be considered indices of aspects of word processing during reading rather than as measures of actual processing time required by words. (Author/FL)