Eighteen dogs undergoing lateral thoracotomy at the left fifth intercostal space were randomly assigned to 1 of 3 postoperative analgesic treatment groups of 6 dogs each as follows: group A, morphine, 1.0 mg/kg of body weight, IM; group B, 0.5% bupivacaine, 1.5 mg/kg given interpleurally; and group C, morphine, 1.0 mg/kg given interpleurally. Heart rate, respiratory rate, arterial blood pressure, arterial blood gas tensions, alveolar-arterial oxygen differences, rectal temperature, pain score, and pulmonary mechanics were recorded hourly for the first 8 hours after surgery, and at postoperative hours 12, 24, and 48. These values were compared with preoperative (control) values for each dog. Serum morphine and cortisol concentrations were measured at 10, 20, and 30 minutes, hours 1 to 8, and 12 hours after treatment administration. All dogs had significant decreases in pHa, PaO2, and oxygen saturation of hemoglobin, and significant increases in PaCO2 and alveolar-arterial oxygen differences in the postoperative period, but these changes were less severe in group-B dogs. Decreases of 50% in lung compliance, and increases of 100 to 200% in work of breathing and of 185 to 383% in pulmonary resistance were observed in all dogs after surgery. Increases in work of breathing were lower, and returned to preoperative values earlier in group-B dogs. The inspiratory time-to-total respiratory time ratio was significantly higher in group-B dogs during postoperative hours 5 to 8, suggesting improved analgesia. Blood pressure was significantly lower in group-A dogs for the postoperative hour. Significant decreases in rectal temperature were observed in all dogs after surgery, and hypothermia was prolonged in dogs of groups A and C. Significant differences in pain score were not observed between treatment groups. Cortisol concentration was high in all dogs after anesthesia and surgery, and was significantly increased in group-B dogs at hours 4 and 8. Significant differences in serum morphine concentration between groups A and C were only observed 10 minutes after treatment administration. In general, significant differences in physiologic variables between groups A and C were not observed. Results of the study indicate that the anesthesia and thoracotomy are associated with significant alterations in pulmonary function and lung mechanics. Interpleurally administered bupivacaine appears to be associated with fewer blood gas alterations and earlier return to normal of certain pulmonary function values. Interpleural administration of morphine does not appear to provide any advantages, in terms of analgesia or pulmonary function, compared with its IM administration.