An experimental model was developed for investigations of traumatically induced vasospasm (TVS) in the central ear artery (CEA) of the awake rabbit. Spasm was induced by a 3-second long pinch with a clip applying forceps exerting a standardized compression force between its claws. The vessel was transilluminated with cold light and inspected from above with an operating microscope equipped with a video camera. The internal diameter was continuously measured. TVS was assessed in terms of its duration, intensity (reduction of initial diameter) and severity (integrated change in diameter over time). The spasm lasted about 10-20 minutes and the inner diameter decreased to about 15-20% of the pretrauma values. The preparation was stable for at least 3 hours and repetitive pinches in different segments of the vessel caused reproducible spasms. TVS was more pronounced when the duration of trauma was prolonged and weakened when the same vessel segment was traumatized repeatedly. The nervous influence was studied after partial denervation of the ear, after partial denervation with additional nervous blockade, and after total denervation achieved by amputation and replantation of the ear. It was found that denervation had no influence on the TVS. Ice cooling was used to decrease the body temperature by 1.0 degree and 1.75 degree C. This caused a significant increase in duration and severity, but not in intensity of TVS. Bleeding by 20% of the total blood volume or hypervolaemic haemodilution with dextran infusion, 20% of total blood volume, did not affect the spasm. Blood flow of CEA was totally obstructed by ligation of the side branches along a 7 cm long segment and subsequent ligation of CEA distal to the segment. This resulted in decreased duration and severity of TVS. Topical application of 2 or 20% lidocaine or of 3% pentobarbital one minute after vessel trauma, when the vessel was maximally constricted, caused prompt resolution of the TVS. The rapid resolution was followed by drug-induced vasoconstriction to about 60% of the initial pre-spasm values. During this vasoconstriction the vessels were refractory to trauma. The results strongly suggest that TVS is a locally elicited phenomenon of vascular smooth muscle, and can be influenced by local application of lidocaine or pentobarbital.