A gas-liquid chromatographic procedure and a high-performance liquid chromatographic (HPLC) procedure for the quantitative determination of the anticonvulsant drug felbamate in serum are presented. Both methods employ methyl felbamate as the internal standard. The efficiencies of felbamate extraction by both C18 solid-phase extraction (SPE) followed by HPLC and liquid-liquid extraction followed by gas chromatography (GC) were evaluated. The mean absolute recovery of felbamate over a serum concentration range of 10-100 mg/L for SPE and liquid-liquid extraction were 89.4 and 84.4%, respectively. Calibration curves for the GC and HPLC procedures were linear from 5 to 200 mg/L felbamate. The overall within-run precision of liquid-liquid extraction followed by GC analysis yielded coefficients of variation Os) of 7.3% at 16 mg/L (n = 10) and 5.8% at 100 mg/L (n = 10). The overall between-run precision for this method calculated by three determinations on a single day for three successive weeks yielded CVs of 7.8% at 16 mg/L (n = 45) and 6.3% at 100 mg/L (n = IO). The overall within-run precision of SPE followed by HPLC analysis yielded a CV of 2.9 % at 50 mg/L (n = 10), and the overall between-run precision for this method calculated by three determinations on a single day for three successive weeks yielded CVs of 3.3% at 25 mg/L (n = 40) and 3.5% at 50 mg/L (n = 37). The SPE-HPLC procedure was found to yield less variable results than the liquid-liquid-GC procedure. Limits of quantitation and detection of both methods were set at 5 mg/L felbamate. Popular anticonvulsant and acid or neutral extractable drugs did not interfere with either the GC or HPLC methods. The results of an interlaboratory, intermethod comparison study performed on patient serum specimens analyzed in one laboratory by the SPE-HPLC procedure and in another laboratory by the liquid-liquid extraction-GC procedure correlated well (r2 = 0.967; n = 41). Both methods provide clinically useful and comparable results.