We have studied ten breast-feeding mothers who were prescribed tricyclic antidepressant drugs and have also compared their infants' development with a similar group (n = 15) who were bottle-fed. Concentrations of tricyclic drugs in maternal plasma and urine, in fore-milk and hind-milk and in infant plasma (n = 6) and urine (n = 9) were measured by gas chromatography (GC) and by an enzyme immunoassay (EIA). The fat concentration in milk was also measured. Infants health and development were monitored by physical examination and by the Bayley Scales of Infant Development up to 30 months. The amounts of tricyclic drugs and their principal metabolites in maternal plasma were significantly correlated with the oral dose and with the amounts in breast-milk. Drug concentrations in fore-milk, but not in hind-milk, increased in line with its fat content, which was maximal in hind-milk. Correlations between gas chromatographic and enzyme immunoassays of maternal samples were high provided that the values for amitriptyline and nortriptyline were excluded; immunoreactivities to these compounds were abnormally high, suggesting that metabolites were also being measured by EIA. The daily doses of drugs ingested by breast-fed infants were about 1% of the maternal dose/kg and the immunoassay detected very small amounts of tricyclics in infants' plasma and urine. No acute toxic effects were found in the ten medicated breast-fed infants and there was no evidence of developmental delays in comparison with bottle-fed infants. Although the number of subjects in this study is small, when taken in conjunction with other published findings, we have not found any reason to prevent mothers who are taking established tricyclic antidepressants from breast-feeding their babies if they want to do so.