We measured plasma and cerebrospinal fluid (CSF) beta-casomorphin-8, a product of beta-casein hydrolysis which has opioid activity, by RIA in women during late pregnancy and lactation and in nonpregnant nonpuerperal women. Before RIA, the samples were acidified and extracted by reverse phase silica gel chromatography, which removed most of the beta-casein. Lactating women had a significantly higher mean plasma beta-casomorphin-8 concentration (2.66 nmol/L; n = 8) than women in late pregnancy (0.82 nmol/L; n = 8) and nonpregnant women (0.32 nmol/L; n = 5). The CSF beta-casomorphin-8 concentration also was significantly higher in lactating women (mean, 0.35 nmol/L; n = 8) than during late pregnancy (0.22 nmol/L; n = 8) or in nonpregnant women (0.15 nmol/L; n = 5). A positive correlation was found between plasma and CSF beta-casomorphin-8 levels in the entire study group. The milk beta-casomorphin-8 concentration, measured in five puerperal women, averaged 19.8 nmol/L. Thus, there is a decreasing concentration gradient between milk and plasma and between plasma and CSF. Chromatographic analysis revealed mol wt heterogeneity of the RIA-active material. In CSF at least three different components were detected, two of mol wt around 900-2,000 and one of approximately 12,000. One of the low mol wt components coeluted in several chromatographic systems with synthetic beta-casomorphin-8 (mol wt, 900). Such a component was not found in milk or plasma, in which the major activity was due to larger sized peptides. The major peaks in milk were around 1,500-2,000 and 12,000 mol wt, corresponding to the larger peaks in CSF. The results suggest that fragments of the milk protein beta-casein may cross the breast parenchyma-blood barrier into plasma and subsequently penetrate the blood-brain barrier to reach the central nervous system. Thus, mammary tissue may assume endocrine function during galactopoiesis, and beta-casein could be considered a prohormone.