1. To test whether the preoptic area controls only non-shivering and the spinal cord only shivering thermogenesis, ten rats were chronically implanted with a preoptic and a spinal cord thermode each. The following were then studied: (a) the effect of propranolol (8 mg/kg.hr) on the metabolic response to cooling the preoptic area, and the spinal cord, (b) the effect of exogenous noradrenaline (0.5 mg/kg) on the metabolic response to cooling the preoptic area, and the spinal cord, and (c) the effect of warming the preoptic area on the metabolic response to cooling the spinal cord, and vice versa. 2. Administration of propranolol inhibited the metabolic response to cooling each of the thermosensitive areas, but the response to cooling the preoptic area was more strongly inhibited than that to cooling the spinal cord. 3. Administration of exogenous noradrenaline did not prevent the metabolic response to cooling either the preoptic area or the spinal cord. 4. Warming the spinal cord completely inhibited the metabolic response to cooling the preoptic area, and warming the preoptic area fully inhibited the metabolic response to cooling the spinal cord. 5. It is concluded that exogenous noradrenaline underestimates the capacity for non-shivering thermogenesis, and that both thermosensitive areas can control both forms of thermogenesis, but that the preoptic area threshold of non-shivering thermogenesis is probably lower than that of shivering, while the spinal cord threshold of shivering is probably lower than that of non-shivering thermogenesis.