Diet data have been used to address a number of theoretical issues. We often calculate the proportion of time spent eating different foods (e.g., fruits, leaves) to place species into dietary categories and contrast morphological or behavioral traits among categories. Yet we have little understanding of how flexible species can be in terms of the plant parts and species consumed. To address this issue, we analyzed data on the diets of red colobus monkeys (Procolobus badius) from Kibale National Park, Uganda, to evaluate temporal and spatial variability in the plant parts and species eaten. After considering observer differences and sampling issues, we evaluated how different a group's diet could be if samples were taken in different years. We found that the diet of the same groups showed significant, consistent changes over a 4-year period. For example, the time spent feeding on leaves increased from 56% in 1994 to 76% in 1998. The plant parts and species eaten by eight groups inhabiting different types of forest (e.g., pristine, logged, riverine) varied among groups. The largest interdemic difference was seen in the use of young leaves (38%). Dietary differences were also found between groups with overlapping home ranges (41–49% overlap). Different subspecies of Procolobus badius also varied in diet; however, this variation was often not of the magnitude documented within Kibale for the same population. The fact that diet can vary considerably over small spatial and short temporal scales within the same species raises the intriguing question as to what level of interspecific difference is biologically significant for addressing particular questions. We conclude that behavioral flexibility blurs our traditional stereotypic assessment of primates; a study of one group that occupies a specific habitat at one point in time may not adequately represent the species. Am J Phys Anthropol 117:349–363, 2002. © 2002 Wiley-Liss, Inc.