SUMMARY The effects of insulin on the concentration of adrenaline and noradrenaline in human plasma and erythrocyte extracts were investigated with the aid of a fluorimetric method of estimation. The following results were obtained: 1. The intramuscular injection of insulin, in doses sufficient to induce hypoglycaemic coma, led to an immediate fall of the adrenaline concentration in plasma. The lowest value was usually found within 0·5 hr of the injection. Subsequent samples may show a slight rise, but the values remained well below the normal level of concentration. Termination of coma was followed by a return of the adrenaline concentration to the starting level or above. The concentration of noradrenaline in plasma and erythrocytes was not affected significantly. The concentration of adrenaline in erythrocytes showed a tendency to decrease. 2. The intravenous injection of insulin (0·1 unit/kg) produced a fall of the level of adrenaline in the plasma which reached its maximum within about 5 min. This was followed by a spontaneous return of the adrenaline concentration to the initial level. The noradrenaline level did not change significantly with the exception of a slight decrease after 90 min. The concentration of both adrenaline and noradrenaline in erythrocytes tended to decrease, but the changes, except for one value, were not statistically significant. 3. The initial change in the concentration of adrenaline in the plasma was independent of the amount of insulin injected, but the higher the dose, the longer the delay in the spontaneous return to the normal level. 4. The termination of hypoglycaemic coma by an intravenous injection of glucose was accompanied by a rapid rise in the concentration of adrenaline in the plasma to a peak value which was reached after 5 min. Later the concentration of adrenaline became stabilized at an intermediate level. The concentration of noradrenaline in the plasma tended to rise, but the change was significant only in a group of samples taken 60 min after the injection of glucose. The termination of coma by a nasal glucose feed produced similar changes, but they were smaller and developed more gradually. 5. An intravenous injection of insulin (0·1 unit/kg) produced a fall in the concentration of adrenaline in venous blood which preceded a similar effect in arterial blood. There was thus a transitory increase in the arterio-venous difference in the concentration of adrenaline. 6. The results are discussed, and it is suggested that they can be explained by assuming that insulin increases the utilization of circulating adrenaline in the phosphorylase reaction.