Some patients with diaphragmatic paralysis or dysfunction maintain ventilation by use of other muscles. Anaesthesia, in modifying the performance of these muscles, presents a potential risk to such patients. To evaluate this risk, the effects of halothane on ventilation and arterial blood gases were studied on a model of bilateral diaphragmatic paralysis, the phrenectomized rat. The study was performed on 43 rats. Success of phrenectomy was confirmed at laparotomy, which did not result in blood gas changes. Laparotomy was performed in 23 rats and a carotid artery was catheterized. In 11 control rats, phrenic nerves were exposed but not sectioned, and in 12 other rats, the phrenic nerves were sectioned. Ventilation was measured by plethysmography in awake rats before and after surgery and in the same rats anaesthetized with halothane 1.1%. In the 23 rats, a decrease in weight and core temperature was observed after operation and this was more marked in phrenectomized than in control rats. In the 11 control rats, ventilation increased postoperatively without change in blood gases. In these rats, halothane caused a decrease in minute ventilation and PaO2 and an increase in PaCO2. Phrenectomy in awake rats led to an increase in minute ventilation, hypoxaemia and hypercapnia. In these rats, halothane led to death in three and a decrease in minute ventilation, with hypercapnia and hypoxaemia in the nine other rats. Blood gas changes were greater than in anaesthetized controls. In the intact rat, halothane leads to blood gas changes comparable to those observed in other species and humans. The present study confirms the effects of halothane on respiratory muscles other than the diaphragm and demonstrates the severe respiratory risk of anaesthesia in patients whose ventilation is maintained by these muscles.