Measurements of blood pressure were obtained on 2,673 women from East Boston, Massachusetts, an urban, working class neighborhood, in surveys conducted in both 1973 and 1976-1977. Of these, 927 women participated in a third survey in 1978. The women were 16 to 49 years of age in 1973, premenopausal throughout the study, and not on blood pressure medication. Regression analyses were performed of blood pressure change between the first and second as well as between the second and third surveys on initial blood pressure, age, weight, and patterns of oral contraceptive use. For systolic pressure the effect of starting oral contraceptive use was an increase of 4.1 mmHg (p less than 0.0001), while the effect of discontinuing use relative to continued use was a 4.4 mmHg decrease (p less than 0.0001). These changes were not affected by duration of use or time since last use among past users. For diastolic pressure the average effect of starting use between surveys was an insignificant 1.0 mmHg increase, but diastolic pressure level seemed to increase with duration of use (0.5 mmHg/year, p = 0.0009). The effect of discontinuing use relative to continued use was a drop of 2.7 mmHg in diastolic pressure (p = 0.0004), which was uninfluenced by time since last use.