Temporal order discrimination in rats was analyzed using a radial maze. The task consisted of a study phase in which two to five items (arms) were presented sequentially and, after a delay, a test phase in which two of these were simultaneously presented and the rat had to choose the arm presented earlier in the study phase. Acquisition of the task was better as the number of items in the study phase increased when the total study phase length was constant (4 min). Interval length between items in the study phase did not affect the acquisition regardless of the number of items. Additionally, discrimination performance was better when two items with more intervening items in the study phase (temporal lag) were used for the test. However, this tendency was obtained only when the last item of the study phase was included in the test pair. Results suggest that the number of items presented in the sequence, but not interval length between items, is an important factor in temporal order memory, and that a larger number of intervening items, as well as containing the last item in the study phase, contributes to the occurrence of the temporal lag effect.