As in our previous report, when cats were fitted with stereotaxic ear-bars 'type A' animals (26 out of 41) still exhibited a parasympathetic reflex lip blood flow (LBF) increase in response to lingual nerve stimulation, while in 'type B' animals (the remaining 15) it was greatly reduced or abolished. We compared (in both magnitude and in their sensitivity to hexamethonium, 10 mg/kg, i.v.) the LBF responses evoked by electrical stimulation of various sites within the reflex arc (lingual nerve, trigeminal ganglion, spinal trigeminal nucleus (Vsp)) in type A and type B animals to examine where the suppressive effect of ear-bar insertion might be exerted (using artificially ventilated, cervically vago-sympathectomized cats deeply anesthetized with alpha-chloralose and urethane). After ear-bar insertion: (a) in type A animals, stimulation of both lingual nerve and Vsp evoked a similar, hexamethonium-sensitive LBF increase; (b) in type B animals (in which lingual-nerve stimulation evoked no LBF increase), Vsp stimulation evoked a hexamethonium-sensitive LBF increase; (c) in both type A and type B animals, trigeminal ganglion stimulation consistently elicited an LBF increase (abolished by hexamethonium in type A, but reduced by only 50% in type B). These results suggest (i) that abolition of the lingual nerve-induced parasympathetic reflex vasodilatation by ear-bar insertion is due to reduced afferent traffic (in peripheral trigeminal or facial nerves) rather than to a damaged efferent output, and (ii) this effect in type B animals seems somehow to allow an antidromic (hexamethonium-insensitive) vasodilatation to occur on trigeminal ganglion stimulation.