The microvascular effects of varying concentrations of lidocaine were evaluated with the use of videomicroscopy in an in vivo rat cremaster muscle preparation. Animals were anesthetized with chloralose and urethane and breathed room air spontaneously. Mean arterial pressure and heart rate were measured via a carotid artery cannula. The cremaster muscle was suffused with a balanced electrolyte solution and pH, temperature, PO2, PCO2, and osmolarity were controlled. Internal diameters of fourth-order arterioles in the cremaster muscle were measured with an electronic vernier system. In one group of animals (n = 7), arteriolar diameters were measured every 30 s during a 10-min control period, a 10-min period of topical application of lidocaine hydrochloride, and a 10-min recovery period. Lidocaine hydrochloride, 10(0), 10(1), 10(2), 10(3), or 10(4) micrograms X ml-1, produced changes in arteriolar diameters to 88.9 +/- 0.9, 79.0 +/- 1.3, 67.5 +/- 2.4, 60.1 +/- 3.4, and 127.1 +/- 7.2 per cent of control, respectively (P less than 0.001). In a second group of animals (n = 4), fourth-order arteriolar diameters were measured during administration of intravenous lidocaine, 1.2 mg X kg-1 bolus plus 0.3 mg X kg-1 X min-1. Vasoconstriction to 91.3 +/- 0.9% of control was observed (P less than 0.001). These results demonstrate a biphasic dose-dependent response to lidocaine. At lesser concentrations, including those that occur in the plasma of patients during intravenous infusion or nerve blocks, dose-related vasoconstriction occurred. Lidocaine, 10(4) micrograms X ml-1, a concentration similar to that which occurs at the site of injection during infiltration, nerve block, or epidural anesthesia, produced vasodilation.(ABSTRACT TRUNCATED AT 250 WORDS)