The entry of prolactin into the cerebrospinal fluid from the blood, and the relation between levels in the two compartments were studied under a variety of conditions in ovariectomized rhesus monkeys. Prolonged treatment with either domperidone or sulpiride, both dopamine-receptor blockers, elevated prolactin levels in serum and cerebrospinal fluid proportionally equally, so that the cerebrospinal fluid serum ratio was unchanged from controls (circa 12-20%). Gel filtration showed that only 'little' (monomeric) prolactin entered the cerebrospinal fluid in such monkeys. Following acute elevations of blood prolactin after a single injection of either sulpiride or ovine prolactin, cerebrospinal fluid levels increased linearly over a 90 min sampling period, despite falling serum levels. The rate of entry of prolactin into the cerebrospinal fluid was similar after either procedure, and was independent of absolute serum or cerebrospinal fluid levels, suggesting a rate-limiting mechanism. Furthermore, retrograde portal blood flow from the pituitary is not necessary to account for these results. Clearance of prolactin from the third ventricle was studied following intraventricular injection of prolactin. Prolactin was removed from the cerebrospinal fluid by a mechanism whose efficiency compares with that in serum, so that the half-life of prolactin in either compartment is about the same. Measurement of sodium, potassium and calcium in the cerebrospinal fluid during prolonged hyperprolactinaemia showed no change, indicating that the central effects of prolactin are not due to alteration of these electrolytes. These results show that cerebrospinal fluid levels of prolactin, and hence, those surrounding the brain, can be inferred accurately from those in the blood, and suggest that there may be a selective mechanism regulating the entry of prolactin into the cerebral compartment.