Changes in the excretion of adenosine 3’:5′-cyclic monophosphate (cyclic AMP) have been reported in depressive illness. Abdulla and Hamadah (1970) reported that urinary cyclic AMP excretion was lower than normal during depression and increased with recovery. However, these results were based on single 24-hour urine collections during depression and on recovery, with no creatinine estimations to suggest that the collections were complete. There was no control of diet, drugs or activity. The controls do not appear to have been matched for age. Paul, Ditzion, Pauk and Janowsky (1970) reported that the cyclic AMP excretion in neurotic depression was higher and in psychotic depression was lower than in a control group, but neither difference was statistically significant. However, on enlarging the study by including more psychotic depressives they reported that the cyclic AMP excretion of this group was significantly less than that of the controls (Paul, Cramer and Goodwin, 1971). These workers had controlled the patients' drug and dietary (but not fluid) intake. There appeared to be only minimal control of activity. The results were based on approximately two samples of urine per subject, which were very carefully checked for completeness of collection. Unfortunately the age of the controls (19–22 years) was very different from that of the patients (25–64 years). On two small groups of patients treated with either Laevodopa or lithium carbonate, they reported that changes in affective state were accompanied by changes in the urinary excretion of cyclic AMP. However, in serial studies on manic-depressive patients Paul, Cramer and Bunney (1971) failed to show a correlation between mood rating and cyclic AMP excretion in five out of seven patients; but they reported that the cyclic AMP excretion was increased on the day of rapid switch from depression to mania. The above groups of workers had used an enzymatic-isotope displacement technique to estimate the cyclic AMP. Brown, Salway, Albano, Hullin and Ekins (1972), using a saturation method to assay cyclic AMP, found no correlation between mood and cyclic AMP excretion in two short-cycle manic-depressive patients. Jenner, Sampson, Thompson, Somerville, Beard and Smith (1972) wrote: ‘We have measured daily excretion by a number of depressed and manic depressive patients over periods covering several mood changes without being able to establish any consistent correlation between cyclic AMP excretion and mood, … However, in one unusual case we have found a very marked correlation‘. We (Naylor, Dick, Dick, Moody and Stansfield, 1974) were unable to demonstrate any relationship between urinary cyclic AMP excretion and mood in a patient with recurrent psychotic episodes, in which depressive features predominated.