Rats rendered chronically epileptic by bilateral implantation of cobalt into frontal cortices were simultaneously prepared with permanent electrodes for longitudinal recording of the electroencephalogram (EMG). Delta-8-tetrahydrocannabinol (Δ-8-THC; 10 mg/kg), delta-9-tetrahydrocannabinol (Δ-9-THC; 10 mg/kg), cannabidiol (CBD; 60 mg/kg), or polyvinylpyrrolidone (PVP) vehicle (2 ml/kg) was administered IP twice daily from day 7 through 10 after cobalt implantation, at which time generalized seizure activity in non-treated cobalt-epileptic rats was maximal. Relative to PVP-treated controls, CBD did not alter the frequency of appearance of seizures during the course of repeated administration. In contrast, both Δ-8-THC and Δ-9-THC markedly reduced the incidence of seizures on the first and second days of administration, Interactal spiking during this period, on the other hand, was actually enhanced. On the third and fourth days, tolerance to the effect on seizures was evident, with a return of seizure frequency of THC-treated rats to values not significantly different from those of controls. Unlike the effect on seizures, no tolerance developed to the marked suppression of rapid eye movement (REM) sleep induces by Δ-8-THC and Δ-9-THC. REM sleep remained reduced in the treated animals during the first 2 days after termination of THC administration. In contrast, REM sleep time was unaffected by repeated administration of CBD. These results suggest that Δ-8-THC and Δ-9-THC exert their initial anticonvulsant effect by limiting the spread of epileptogenic activity originating from the cobalt focus.