A 12-week randomized experiment investigated the effects of daily relaxation breaks on office workers with "normal" blood pressure. Blood pressures of 126 volunteers from the corporate offices of a manufacturing firm were measured biweekly. After four weeks of baseline monitoring,volunteers were divided randomly into three groups: Group A was taught a technique for producing the relaxation response; Group B was "taught" to sit quietly; and Group C was taught nothing. Groups A and B were asked to take two 15-minute relaxation breaks daily. During the baseline period, mean systolic blood pressures (SBP) were 119.7, 118.4, and 114.2 for Groups A, B and C respectively; mean diastolic pressures (DBP) were 78.7 and 75.7 Between the first and last measurements, mean changes in SBP were -11.6, -6.5, and +0.4 mm Hg in Groups A, B, and C; mean DBP decreased by 7.9, 3.1, and 0.3. Between the fourweek baseline period and last four weeks of the experimental period, mean SBP and DBP, mean changes in Group A were significantly greater than those in Group B (p < 0.05) and in Group C (p < 0.001). The same pattern of changes among the three groups was exhibited by both sexes, all ages, and at all initial levels of blood pressure. However, in general, within Group A, the higher the initial blood pressure, the greater the decrease. (Am. J. Public Health 67;954-959, 1977)