Following kainic acid (KA)-induced status epilepticus (SE), the ketogenic diet (KD) retards the development of epileptogenesis, with fewer spontaneous recurrent seizures (SRS) and less mossy fiber sprouting than rats on a normal diet. In this study, we investigated whether there is a critical period for initiation of the KD, in terms of the diet's effectiveness in reducing SRS. In addition, we investigated whether early treatment with the KD prevents the deficits in spatial learning and memory that ordinarily follow KA-induced SE. Young rats (P30) underwent KA-induced SE, followed by assignment to one of three treatment groups: control diet ('KA'), KD begun 2 days after SE ('KD2'), and KD begun fourteen days after SE ('KD14'). For 12 weeks following SE, rats were monitored by closed circuit video recording (12 h/wk) to detect SRS. KD2 rats had significantly fewer SRS than rats in the control or KD14 groups. On water maze testing to assess spatial learning and memory, KD2 rats had significantly poorer acquisition of place learning than control (KA alone) or KD14 rats. KD2 rats also failed to gain weight well. There was no difference between groups on routine histologic examination of the hippocampus. In summary, P30 rats placed on the KD 2 days after SE were relatively protected from recurrent seizures, but showed behavioral and physical impairment. Rats placed on the KD 14 days after KA-induced SE did not differ from controls with regard to spontaneous seizure rate.