Investigations into the possible causes of toxic symptoms arising during the course of prolonged somnifaine narcosis showed that a fall in serum calcium occurred during the course of the treatment (1). Previous investigations on blood calcium (3, 4, 5, 6, 15) made it clear that during barbiturate narcosis there was a fall of 10–20% of the normal value. Observations, however (7, 8, 9, 10), that therapeutic exhibition of insulin resulted in an increase in serum calcium suggested that no gross abnormality in calcium level was to be expected during the modified somnifaine-insulin-glucose treatment (2) practised in this hospital. The investigations here reported were undertaken to determine what fall in serum calcium occurred during this modified treatment, and to determine whether this fall was sufficient to cause toxic symptoms. If so the advisability of taking steps to avoid them could then be considered. The investigations showed a definite lowering in serum calcium during the course of treatment of psychotic patients by this method, and they suggested that somnifaine and insulin are only pseudo-antagonistic in their effect upon calcium metabolism. In the ten cases reported, a fall in serum calcium of from 3–20% has invariably been observed, the maximum being reached four to seven days after the beginning of the treatment.