Levels of in serum and cerebrospinal fluid have been studied in male and female rhesus monkeys. Untreated animals showed circadian rhythms of cortisol in both compartments, highest values occurring at 08.00 h and lowest at 20.00 h, but the decline following the morning peak was slower in cerebrospinal fluid (CSF) than in serum. Both levels and CSF/serum ratios (c. 0.017-0.027) were similar in males and females. The ratio between highest and lowest points on the circadian rhythm was greater for CSF (males: 2.50; females: 2.63) than for serum (1.69 and 1.78). There were significant correlations between CSF and serum levels in individual monkeys at each of the four time points studied. No circadian rhythm in CSF homovanillic acid (HVA) or 5-hydroxyindoleacetic acid (5-HIAA) was found, nor any correlation between these monoamine metabolites and cortisol levels. Injecting cortisol once daily in the evening (18.00h) resulted in greater proportional elevation in the CSF than in serum cortisol in samples taken two hours later, and the values in the two compartments were no longer correlated at this time. There was no effect on CSF 5-HIAA levels, Injecting L-5-hydroxytryptophan (L-5-HTP) (2.5 mg/kg) at four different time points produced abnormal daily patterns of cortisol secretion, the peak following the injection of L-5-HTP. The distortion of the normal circadian rhythm thus induced was greater in the CSF than in he serum, and CSF/serum ratios were increased one hour following L-5-HTP administration. The data suggested, but did not prove, that L-5-HTP injection may have produced the least Proportional elevation of cortisol at 08.00 h (at the time of normal daily maxima). 5-HIAA, but not HVA, was elevated one hour after L-5-HTP to 2-3 ties normal . A dose of 0.2 mg dexamethasone at 23.00 h suppressed cortisol levels in serum examined at 08.00 h the next day. This suppression was partially reversed by giving L-5-HTP one hour before taking the next morning's sample. Rhythmic and episodic changes in serum cortisol thus have a proportionally greater effect on the daily pattern of cortisol level in the CSF (and hence the cerebral compartment) than that in the vascular compartment, and thus the exposure of cortisol-sensitive tissues in the two compartments to cortisol may differ.