Regional intravascular sympathetic (RIS) block was produced in the rabbit ear by intra-arterial injection using guanethidine, during tourniquet occlusion of the ear vessels. Dose-effect relations were studied for the fall in mean arterial blood pressure after tourniquet release 30 seconds and 10 minutes after guanethidine administration. Compared to systemic intravenous injection, the effects after 30 seconds' occlusion were reduced by 64% and after 10 minutes' occlusion by 80%, irrespective of the dose used. This indicates a rapid and a-specific 'binding' of guanethidine in the tissues. After a RIS block (10 minutes' occlusion), vasoconstriction during a cold provocation test was abolished. Skin temperature in the treated ear remained high while the temperature of the untreated contralateral ear fell sharply. The block lasted 3 days at 0.625 mg and 5 days at 2.5 and 10 mg. A second block applied within 5 days was less effective, but applied after three weeks it was as effective as the first block. If the results obtained may be applied to man, it may be concluded that RIS block is a relatively safe method, since the major part of the administered dose will rapidly become unavailable for an immediate systemic effect.