To standardise the method of eliciting somatosensory-evoked potentials (SEPs), SEPs were generated by electrical stimulation of different stimulus sites and recorded bilaterally from the primary somatosensory cortex (S1) and from midline in awake, freely moving rats. Increasing stimulus intensity enhanced amplitudes of all SEPs. At supramaximal stimulation, SEPs following vibrissae and tail stimulation (V-SEP and Ta-SEP, respectively) but not following trunk stimulation (Tr-SEP), fulfilled our criterion of signal-to-noise ratio >or=4. The first V-SEP component coincided with a stimulus artefact, disqualifying these recordings for a standard stimulation protocol. The Ta-SEP generated stable and reproducible recordings and was considered to be the preferred technique. Early components of the contralateral S1 recorded V-SEP and Tr-SEP occurred at latencies different from the other recordings. Increasing stimulus repetition rate (SRR) decreased amplitudes of all SEPs. At the highest obtainable SRR, the amplitude between the V-SEP second positive and second negative components in all recordings was 70-80% of the amplitude at 0.1 Hz, whereas peak amplitudes of subsequent components and those of the Tr-SEP and Ta-SEP were 20-50%. These results indicate that the different SEP components might be generated by different ascending neural pathways.