One hundred and eighty three hypertensive pregnant women were randomly assigned to antihypertensive treatment with oxprenolol (96 women) or methyldopa (87 women). Control of hypertension was equivalent in both treatment groups, and in 64 (35%) cases hydralazine had to be added to the treatment to achieve the therapeutic goal (diastolic blood pressure below 85 mm Hg). Five perinatal deaths occurred, one in the oxprenolol group and four in the methyldopa group. Detailed analysis confirmed a previous report of greater fetal growth in the group treated with oxprenolol; this trend was present regardless of severity of hypertension and parity. With increasing duration of treatment the differences between the two groups diminished, and there was no difference after 10 weeks of treatment, a finding that may explain some of the reported discrepancies among therapeutic studies. As hypertension in pregnancy may pursue an accelerated course, necessitating urgent delivery, and there is no satisfactory method of predicting the duration of treatment in individual patients fetal benefit is most likely to be achieved by treatment with oxprenolol, provided that there is no maternal contraindication to treatment with beta blockers.