The relative CNS and cardiovascular toxicity of bupivacaine was compared in pregnant and nonpregnant ewes during continuous infusion of bupivacaine into the jugular vein at the rate of 0.5 mg .cntdot. kg-1 .cntdot. min-1. In all animals, identical symptoms of toxicity occurred in the following order: convulsions, hypotension, respiratory arrest and circulatory collapse. The dose of bupivacaine required to produce CNS toxicity in the pregnant ewe tended to be lower than in the nonpregnant animal, although the difference was not statistically significant (P < 0.1). The mean dose of bupivacaine resulting in cardiovascular collapse was significantly lower in pregnant ewes (5.1 .+-. 0.7 mg/kg) than in nonpregnant animals (8.9 .+-. 0.9 mg/kg). Similarly, bupivacaine blood concentrations at the onset of respiratory arrest and circulatory collapse were lower in the pregnant group, being 5.2 .+-. 0.7 .mu.g/ml and 5.5 .+-. 0.8 .mu.g/ml, respectively, vs. 7.5 .+-. 1.0 .mu.g/ml and 8.0 .+-. 0.9 .mu.g/ml, respectively, in the nonpregnant group (P < 0.05). The concentration of bupivacaine in the brain of pregnant ewes at the time of cardiovascular collapse was significantly lower (P < 0.01) than in the nonpregnant group (7.5 .+-. 1.5 vs. 16.3 .+-. 1.7 .mu.g/g). The myocardial tissue concentration of bupivacaine also tended to be lower in the pregnant group, although the differences were not statistically significant (P < 0.1). Comparisons of bupivacaine doses and blood concentrations associated with the onset of convulsions and circulatory collapse (CC/CNS ratio) with those of lidocaine, reported previously, indicate that a narrower margin of safety exists following administration of bupivacaine in nonpregnant sheep. These ratios were 3.7 .+-. 0.5 and 1.6 .+-. 0.1, respectively, for bupivacaine and 7.1 .+-. 1.1 and 3.6 .+-. 0.3, respectively, for lidocaine. The pregnant sheep may be more sensitive to the cardiotoxic effects of bupivacaine than the nonpregnant animal.