In previous investigations of changes observed in the blood levels of acetylmethylcarbinol and butane-2: 3-diol in patients with manic-depressive psychosis (Dawson et al., 1954), the blood levels of a large number of patients in one mood phase were compared with the levels in another group of patients in a different mood phase. This procedure showed that the differences between the various mood phases were statistically significant. A lack of correlation in time between the chemical changes and the alteration in mood might account for the limited degree of correlation obtained. Thus, if the chemical change preceded the mood change, some of the data would not correspond with the observed mood state during which the blood sample was taken.

Another uncontrolled factor in the previous investigation was the diet; whilst all the blood samples were taken after a 12 hour fast, there was a large difference in the previous dietary history of the patients due to their mental state. Again, only blood levels were investigated and these might have been considerably affected by alterations in kidney activity.

For all these reasons, a different approach to the problem was made, and a small number of patients has been studied closely through a series of mood changes so as to enable us to determine the time relationship, if any, between the change in mental state and the chemical change. The patients were segregated into a small ward where their food intake was determined and 24-hour urine collections made; whilst in the unit, patients were maintained without sedation and without electroconvulsant therapy.

On the basis of the previous work, it was concluded that the variations observed in the blood levels of acetylmethylcarbinol and butane-2: 3-diol did not, in themselves, determine a change in mood but were indicators of some other, more fundamental alteration in the metabolic state of the subject. Consequently, changes in extracellular fluid volume as estimated by determinations of the thiocyanate space, were also investigated with these patients. Estimations of urinary constituents were also made to determine the effect of the kidneys in influencing blood levels.