In vitro studies have demonstrated that neuroeffector transmission is potentiated when stimulating the innervation of the caudal ventral artery with short bursts of high frequency spike trains, rather than continuous trains of impulses with the same frequency. Furthermore, in anaesthetised CNS-intact rats, sympathetic neuronal activity regulating the tail vasculature (SNAT) tends to occur in bursts which have a robust rhythmicity in the 0.4-1.2 Hz frequency range (T-rhythm). These rhythmic discharges appear to be generated by autonomous sympathetic oscillators (Gilbey, 2001). Rostral ventromedial medullary neurons are critically involved in regulating SNAT (Korsak & Gilbey, 2004) and some contain serotonin and/or v-glut3 transporter (Smith et al. 1998; Nakamura et al. 2004). In the present study, we hypothesized that the so-called T rhythm is generated within the spinal cord, and dependent upon serotonergic and glutamatergic inputs. Population sympathetic activity was recorded from the dorsal collector nerve (DCN) of the tail in urethane anaesthetised (1.3 g kg^-1, i.p., supplemented with 5-10 mg i.v. as required) rats spinalized between T10 and T11. Intrathecal (i.t.) injections (10 μl over 5 min) were delivered to L1, as sympathetic preganglionic neurons projecting to the DCN are distributed in T11-L2 (Smith & Gilbey, 1998). Nerve activity was completely abolished after spinal transection (n=15) and no spontaneous recovery was observed afterwards (n=3). The administration of 5-HT (250 nmol, i.t.) induced sympathetic rhythmic discharges (median frequency, 0.90 Hz; interquartile interval, 0.84-1.01 Hz; n=6) that were sustained for up to 90 min and were similar to those observed before spinal cord transection (median frequency, 0.77 Hz; interquartile interval, 0.72-0.83 Hz; n=6). In contrast, NMDA (1 μmol, i.t.) failed to restore the T-rhythm, but induced activity that was sustained for long periods of time (at least 4 h). This tonic activity induced by NMDA became rhythmical for periods of 30-60 min when 5-HT (250 nmol, i.t.) was additionally administered (median frequency, 0.93 Hz; interquartile interval, 0.70-0.96 Hz; n=6). In all cases, it was possible to repeatedly initiate sequences of stable rhythmic activity by performing reapplications of 5-HT. Remarkably, the level of activity was similar in all conditions. Our observations are consistent with the idea that serotonin, released from bulbospinal neurones, acts on neurones within the spinal cord to generate the T-rhythm.