We examined serial pattern retention in rats focusing on retention for a single pattern element that violated pattern structure. Violation elements were targeted because they are typically unusually difficult to learn, contrasting with rule-learning in the structured pattern. They are also sensitive to selective manipulation of central neurotransmitter and plasticity systems. Adult male and female rats were trained in a serial multiple choice task to perform nosepoke responses for water in receptacles mounted on the 8 walls of an octagonal chamber. Rats learned to nosepoke in the pattern, 123-234-345-456-567-678-781-818, where digits represent the clockwise position of successive correct receptacles, dashes indicate brief pauses, and the final “violation element” is inconsistent with pattern structure. Rats were trained twice to an 85% correct criterion on the violation element prior to testing after 4-week and 2-week retention intervals, respectively. After each retention interval, rats received a 1-day 10-pattern retention test on the same pattern. Results indicated that rats had approximately 57% retention of the violation element after 2 weeks and 28% retention after 4 weeks. Sex differences were not significant. The results indicate that serial pattern memory, particularly violation element memory, can be assessed by characterizing forgetting curves using traditional methods.