1. Repeated carotid sinus nerve (CSN) stimulation evokes a serotonin-dependent long-term facilitation (LTF) of phrenic nerve activity in cats. To determine whether CSN stimulation-evoked LTF is a general property of spinal inspiratory motoneurones, phrenic and inspiratory internal intercostal (IIC) nerve activities were recorded in nine cats (eight anaesthetized; one decerebrate), which were vagotomized, paralysed, thoracotomized and ventilated with O2; airway CO2 was controlled by means of of a servo-respirator. Baseline conditions were established by setting the arterial CO2 pressure (Pa,CO2) at approximately 2 mmHg above the threshold for IIC activity. One CSN was stimulated (3 times threshold, 25 Hz, 0.5 ms duration) with five (2 min) trains, each separated by 5 min. 2. The peak integrated phrenic activity was elevated by 33% whereas IIC activity was elevated by 226% above baseline, 90 min post-stimulation (P < 0.05). The results were similar when expressed as a percentage of the maximal neural activities (elicited by combined hypercapnia and CSN stimulation), although differences between the nerves were less pronounced. The burst frequency was not change following stimulation. 3. In five additional cats that were pretreated with the serotonin receptor antagonist, methysergide maleate (0.5-1 mg kg-1, I.V.), the CO2 thresholds of the phrenic (12 mmHg) and IIC nerves (22 mmHg) were increased (P < 0.05), and LTF could not be elicited in either neurogram. 4. Successive CSN stimulation episodes evoked a previously undescribed phenomenon. Although the peak integrated phrenic activity was unchanged (90-95% of maximal), IIC activity increased progressively during successive stimulus episodes (66-90% of maximal; P < 0.05). However, after methysergide treatment, the initial stimulus-evoked phrenic response decreased to 58% of maximal and both neurograms exhibited progressive augmentation of the stimulus-evoked response. As stimulus-evoked augmentation does not require serotonin, it is independent of LTF. 5. We conclude that CSN stimulation-evoked LTF of IIC activity exceeds that of phrenic activity. Since LTF requires the neuromodulator serotonin and is expressed predominantly by changes in burst pattern formation versus rhythm generation, serotonin may exert a greater influence on IIC relative to phrenic respiratory motor output. A unique mechanism is described whereby successive CSN stimulus episodes cause progressively increasing responses in both neurograms.