Current biological monitoring techniques are often unable to provide accurate estimates of pesticide dose in exposed worker populations. This study was conducted to investigate the feasibility of pesticide biomonitoring using saliva. Atrazine [2-chloro-4-ethylamino-6-(isopropylamino)-s-triazine], a member of the triazine herbicides, was selected to investigate salivary excretion following direct gastric administration in rats. Concentrations of atrazine in whole saliva and arterial plasma samples were determined by enzyme-linked immunosorbent assay (ELISA). Atrazine reached its highest level in both arterial plasma (238 g/L) and whole saliva (157 g/L) 35 min after administration of 105 mg/kg of atrazine, and then decreased with time in a parallel fashion. Although saliva atrazine levels were lower than levels in arterial plasma, there was a very high correlation between whole saliva and 2 arterial plasma atrazine concentrations (r = .95). In addition, pharmacokinetic analysis suggested that salivary levels of atrazine can be used to predict concentrations of atrazine in plasma. The mean whole saliva/arterial plasma atrazine concentration ratio (S/P) was 0.66 0.11 (n = 20). The S/P ratios did not vary significantly over time, and were not affected by salivary flow rate. This study demonstrates that atrazine is transported into saliva, and that a relatively constant concentration ratio between whole saliva and arterial plasma is maintained. Because the salivary concentrations of atrazine are independent of variation in salivary flow rate, salivary monitoring of atrazine in humans may prove useful and practical. Finally, this study suggests that other pesticides with chemical and physical properties similar to those of atrazine can be monitored in saliva.