SUMMARY The serum half‐life of primidone was determined in 6 subjects and was found to be 10 to 12 hours. In 2 subjects who were taking phenobarbital, no difference in the absorption curve and serum half‐life of primidone was noted. This suggests that phenobarbital does not influence the half‐life of primidone. No detectable phenobarbital was present in the serum of the volunteer subjects following the single 500 mg dose of primidone, even up to 48 hours. This suggests that the bioconversion of primidone to phenobarbital occurs slowly. On the other hand, clinic subjects who were taking primidone chronically with no barbiturate showed a mean phenobarbital level 3 times higher than the primidone level. This could be explained by an increasing rate of biotransformation of primidone to phenobarbital following chronic administration. However, a low transformation rate and the marked difference in the serum half‐life of phenobarbital and primidone could also explain the results. Toxic side effects similar to those attributed to phenobarbital and diphenylhydantoin were seen in the volunteer subjects even though these drugs were not present in their serum. In addition, data from a small group of clinic patients suggest that ataxia, somnolence and lethargy could be attributed to high primidone levels. Data relating the mean and range of primidone levels to daily dose are presented as a guide for evaluating the status of an individual patient. A significant increase in blood level with increasing dose was present. Some data relevant to the therapeutic serum level of primidone are presented. They suggest that a serum level below 10–12 μg/ml should be maintained, as the risk of side effects increases greatly above this level. There was no evidence from the present study to suggest that higher levels would be associated with an increase in seizure control. RÉSUMÉ Chez 6 sujets on a trouvé que la demi‐vie de la primidone était de 10 à 12 heures. Chez 2 sujets qui prenaient aussi du phénobarbital on n'a retrouvé aucune différence de la courbe 'absorption et de la demi‐vie de la primidone. Ches des sujets volontaires qui avaient pris une seule dose de 500 mg de primidone on ne retrouvait pas de phénobarbital dans le sérum, même 48 h après. Ceci suggère que la transformation de la primidone en phénobarbital se fait lentement. 'autre part, les sujets qui prenaient de la primidone 'une façon chronique, sans barbituriques, avaient des taux moyens de phénobarbital trois fois plus élevés que le taux de primidone. Ceci peut s'expliquer par une augmentation de la vitesse de transformation de la primidone en phénobarbital à la suite du traitement chronique. Toutefois, les résultats peuvent s'expliquer aussi par une lente vitesse de transformation et par 'importante différence entre la demi‐vie dans le sérum du phénobarbital et celle de la primidone. On a observé des effets toxiques secondaires semblables à ceux qu'on dit être consécutifs au phénobarbital et aux hydantoïnes chez les sujets volontaires, bien que ces produits ne fussent pas...