1. Six lactating white women, aged 24-36 years, received a single oral dose of 300 mg moclobemide, between 09.00 h and 11.00 h, 3 to 5 days after the delivery of a full term neonate. 2. Complete milk collections were obtained before, 3, 6, 9, 12 and 24 h after drug administration by means of a breast pump. Venous blood samples were drawn before, and 0.5, 1, 3, 4.5, 6, 9, 12, 24 h post-dosing. 3. Moclobemide, and its major metabolite (Ro 12-8095) were measured in milk and plasma samples using h.p.l.c. The active metabolite (Ro 12-5637) could only be detected in plasma. 4. Moclobemide and its metabolites were not detectable in 24 h plasma samples. Cmax, tmax and t1/2 for moclobemide were (mean +/- s.d.) 2.70 +/- 1.24 mg l-1, 2.03 +/- 1.19 h and 2.26 +/- 0.26 h, respectively. 5. The concentrations of moclobemide and Ro 12-8095 in milk were highest at 3 h after drug administration and the drug and metabolite were not detectable after 12 h. Ro 12-5637 was not detected in any milk sample. The percentages of the dose excreted as moclobemide and Ro 12-8095 were (mean +/- s.d.) 0.057 +/- 0.020% and 0.031 +/- 0.011%, respectively. An average 3.5 kg breast-fed neonate would therefore be exposed to only a 0.05 mg kg-1 moclobemide dose (approximately 1% of the maternal dose on the mg kg-1 basis). The low amount of moclobemide excreted into breast milk is unlikely to be hazardous to suckling infants.