This is a study of 4 thyrotoxic patients in whom paralysis developed spontaneously shortly after their admission to the hospital and 12 thyrotoxic patients who had a recent history of periodic paralysis. An attempt was made in the latter patients to induce paralysis with a high carbohydrate diet; it was successful in all. For comparison, a similar attempt was made in 6 healthy euthyroid control subjects and 6 thyrotoxic patients with no history of paralysis. No paralysis resulted. Two patients who had spontaneous periodic paralysis and five patients in whom paralysis was induced were given propranolol for 6 days. Spontaneous paralysis did not occur during this period. On the last day induction was attempted. Paralysis failed to occur in four, was greatly reduced in one but occurred with the same intensity in the two patients with observed spontaneous paralysis. Thus propranolol gave partial or complete protection from induced paralysis in five of seven patients (71 per cent). When five of these seven patients were given acetazolamide, they all noticed some weakness of the lower limbs and spontaneous paralysis developed in two. When induction was attempted in four patients on the 6th day of acetazolamide administration, the resulting paralysis was more severe than that in the previous induction. Estimations of plasma insulin levels during the course of induction and during spontaneous or induced attacks failed to confirm the observation of Shishiba and his co-workers (Metabolism 21: 285, 1972) that plasma insulin levels invariably rose to high levels on such occasions. The response to propranolol could not be explained by a depression of the insulin response to high carbohydrate feeding.