1. The contribution of cardiac vagal C fibres to vagal chronotropic control in anaesthetized cats, rats and rabbits was analysed using electrical stimulation of the vagus nerve with a selective anodal block technique. 2. After bilateral vagotomy and pretreatment with atenolol, 10 Hz continuous selective stimulation of unmyelinated fibres in the cut peripheral end of the cervical vagus evoked a bradyeardia in anaesthetized rats, cats and rabbits. With this stimulation protocol the three species exhibited a similar lengthening of the heart period (R-R interval) when expressed as a percentage of their basal cardiac interval. 3. The mechanism of action of the selective blocking technique was analysed by recording eighty-nine single A-(n = 12), B-(n = 22) and C-fibre (n = 55) vagal-projecting neurones in the medulla of the rat. This demonstrated that the technique can selectively block conduction in myelinated fibres and that 'break excitation' is seen mainly in unmyelinated fibres. Although thirty C fibres showed break excitation sixteen did not and this difference could not be correlated with their axonal conduction velocity, chronaxie or initial segment frequency following. 4. Using the anodal block technique the vagal effects on heart rate were reanalysed in the cat by incorporating a collision technique. B fibres were activated orthodromically to evoke cardioinhibition and simultaneously antidromically to collide with errant B-fibre spikes activated at the electrode producing anodal block. With this protocol it was noted that the B-and C-fibre bradycardias were not additive. Using a double anodal block and collision technique, it was demonstrated that this phenomenon was likely to be due to occlusion of the effects of B and C fibres. 5. In conclusion, in addition to the well-defined effects of vagal B fibres on heart rate, selective stimulation of vagal C fibres also had a cardioinhibitory effect in all three species studied. However, since the effects of cardiac C fibres on heart rate was small, these neurones alone cannot account for the cardioinhibition of the pulmonary chemoreflex. It is likely that activation of both B-and C-fibre cardiac vagal preganglionic neurones accounts for this reflex cardioinhibition.