Interictal and ictal discharges are recorded from limbic structures in temporal lobe epilepsy patients. In clinical practice, interictal spikes are used to localize the epileptogenic area, but they also are assumed to promote ictal events. Here I review data obtained from combined slices of mouse hippocampus-entorhinal cortex that indicate an inverse relation between interictal and ictal events. In this preparation, application of 4-aminopyridine or Mg2+-free medium induce (a) interictal discharges that originated from CA3 and propagate (via the Schaffer collaterals) to CA1 and entorhinal cortex, to return to the hippocampus through the dentate area; and (b) ictal discharges that initiate in the entorhinal cortex and propagate to the hippocampus via the dentate gyrus. Interictal activity occurs throughout the experiment (up to 6 h), whereas ictal discharges disappear after 1-2 h. Schaffer collateral cut abolishes interictal discharges in CA1, entorhinal cortex, and dentate and reestablishes entorhinal ictal discharges. Moreover, ictal discharge generation in the entorhinal cortex after Schaffer collateral cut is prevented by mimicking CA3 activity with rhythmic electrical stimulation of CA1 outputs. Thus hippocampal interictal activity controls the ability of the entorhinal cortex to generate seizures. It also may be proposed that Schaffer collateral cut may model the epileptic condition in which CA3 damage results in loss of hippocampal control over the entorhinal cortex. In conclusion, these experiments demonstrate that interictal activity controls rather than promotes ictal events, and functional integrity of CA3 constitutes a critical control mechanism in temporal lobe epilepsy.