Six healthy, female, mixed-breed 18-24-month-old sheep weighing 30-48 kg were submitted to lumbosacral epidural bupivacaine in combination with either methadone or fentanyl. Epidural catheters were placed in six sheep that were given three treatments: (Bup) bupivacaine (0.5 mg/kg) alone; (BupMet) bupivacaine (0.25 mg/kg) plus methadone (0.3 mg/kg); and (BupFent) bupivacaine (0.25 mg/kg) plus fentanyl (0.002 mg/kg). Haemodynamic variables, respiratory rate, rectal temperature, analgesia by applying a standard painful stimulus, motor block and sedative scores were compared among the three treatments. These parameters were determined before epidural administration and at 5, 10, 20, 30, 60, 90, 120 minutes after treatment administration, and then every 60 minutes thereafter until the end of analgesic effect. Parametrical data were analysed by proc glimmix (SAS) for repeated measures on time and means tested by ls-means. Non-parametrical data were analysed by Fisher's exact test. Duration of analgesia was longer with BupMet (240 minutes) compared with BupFent (180 minutes; P=0.0127), but BupMet was similar to Bup (240 minutes). Both treatments with opioids produced moderate motor blockade. BupMet and BupFent produced mild sedation. Only treatment with bupivacaine alone induced cardiovascular and respiratory rate changes that stayed within acceptable limits.