Kurz et al. conducted the first study of the intra-individual variability of clozapine plasma concentrations but did not take into account the effect of smoking and co-medication. As patients were receiving varying doses, Kurz et al. standardized plasma levels by using a plasma level/dose/kg ratio. In 15 patients, the mean coefficient of variation (CV) was 53% (S.D. = 21). In this new study, plasma clozapine and norclozapine concentrations were measured every 2 weeks in 47 patients randomized to 100, 300, or 600 mg/day for 16-week double-blind clozapine trials under controlled conditions (stable smoking, limited co-medication and absence of caffeinated beverages). For 100, 300 and 600 mg/day, the respective mean CVs for plasma clozapine concentrations were 23% (S.D. = 14), 19% (S.D.= 11) and 18% (S.D. = 8). For the combined concentrations of clozapine and norclozapine, the respective mean CVs were 20% (S.D. = 13), 16% (S.D. = 9) and 15% (S.D. = 7). Under 100 mg/day, the mean CV for clozapine concentrations was significantly higher for heavy smokers than non-heavy smokers (32%, S.D. = 3 vs. 19%, S.D. = 8) (p = 0.03). Studies of CVs in other environments are needed. Clozapine CVs may be important in order to understand the importance of variations around the therapeutic range and to interpret drug interactions above the usual noise of measuring plasma concentrations.