Our objectives were to determine whether magnesium sulfate crosses the blood-brain barrier and whether it has central anticonvulsant action. In experiment 1 34 female Long-Evans rats were divided into six groups: control (n = 7); single magnesium sulfate injection and evaluation after 20 minutes in 3 conditions: normal rats (n = 7), sham-operated animals (n = 5), and after electrical stimulation by hippocampal electrode (n = 5); single injection and evaluation after 2 hours (n = 5); and prolonged (2 hours) serum magnesium elevation (n = 5). Serum, cerebrospinal fluid, and specific brain areas were analyzed for magnesium concentrations. In experiment 2 threshold for electrical seizure was measured in eight rats before and after intraperitoneal injections of magnesium sulfate versus saline solution. In experiment 1 there was a significant correlation between blood and cerebrospinal fluid magnesium concentrations (r = 0.80, p < 0.0001). Magnesium concentrations were increased in the cortex and hippocampus, with the largest changes occurring after two hours of sustained serum magnesium concentrations (p < 0.01). Induction of hippocampal seizure activity resulted in further elevations in cerebrospinal fluid magnesium concentrations but did not change brain concentrations. In experiment 2 magnesium sulfate increased the electrical threshold required to induce seizures by 34% (p = 0.01). Magnesium sulfate enters the cerebrospinal fluid and brain and has a central anticonvulsant effect.