The effect of Mg status on the behavior of rats, as determined by nonspecific excitability level and audio genic seizure susceptibility, was investigated. Also, selected biochemical and neurological mechanisms mediating the chain of events from dietary magnesium deficit to the hyper excitability symptoms were examined. Weanling rats fed a low magnesium (10 ppm) diet for 14 days had reduced serum, cerebrospinal fluid and brain magnesium concentrations, and increased brain (Na+ K+)-ATPase activity. They exhibited increased NEL and became highly susceptible to audio genic seizures. Through dietary manipulation and intraperitoneal and cerebral intraventricular injections, it was possible selectively to alter either serum or cerebrospinal fluid magnesium concentrations. Both nonspecific excitability level and audio genic seizure susceptibility responded inversely to cerebrospinal fluid magnesium concentration, but not to serum magnesium, unless the latter was very high. In this instance, nonspecific excitability level but not seizure activity was depressed. Brain serotonin concentration was elevated in 150- 200 g rats fed a low magnesium diet for 21 days, compared with rats fed a control diet (0.73 μg/g fresh brain cf. 0.56 μg/g) . The magnesium deficient rat may serve as an excellent model for epilepsy research. The animal can often be revived following seizure, thus enabling the study of drug interactions while using a small number of subjects. Preliminary findings indicate that the most promising subjects are female rats between 3 and 5 weeks old at the start of feeding and who are fed a low magnesium diet 17 to 21 days. A procedure for determining (Na+ K+)-ATPase activity in rat brain homogenates, without requiring isolation or purification, is described. Computer modeling techniques have yielded expressions for equating enzyme activity to sodium, potassium, magnesium or calcium concentrations in the reaction media. A thorough statistical treatment of the data is presented.