Blood pressure variability was studied in 8 "labile" and in 8 "established" hypertensive patients by means of continuous unrestricted blood pressure recording for 24 hours. In 7 of the former and 5 of the latter the effect of atenolol (200 mg/day) was also investigated. Blood pressure variability as judged by circadian rhythm, standard deviation, coefficient of variation, indexes of skewness and kurtosis was similar in the two groups. The only significant difference consisted in the mean arterial pressure which was higher in the established than in the labile hypertensives. Atenolol produced in all subjects a fall in blood pressure but no significant changes in the standard deviation, coefficient of variation and index of skewness. On the contrary the index of kurtosis significantly increased suggesting a more stable blood pressure after treatment.