Controversy exists over the value of measuring diastolic blood pressure (BP) in children, and over whether this should be measured at Korotkoff's fourth phase (K4) or fifth phase (K5) of pulse sounds. We measured diastolic BP in 3012 randomly selected Finnish children aged 6-18 years in 1980, and in 2885 of the same individuals in 1983 and 2500 in 1986. BP was measured with a standard mercury sphygmomanometer in 1980 and 1983, and with a random-zero sphygmomanometer in 1986. K4 was consistently absent in 187 individuals (3.2%) in 1980 and 1983, and in 155 individuals (6.2%) in 1986. K5 was absent in only 34 individuals (0.6%) in 1980/83 and 5 (0.2%) in 1986. The difference between mean K4 and K5 diastolic BP varied from 6.5 to 9.2 mm Hg depending upon age. K4 and K5 BPs showed good correlation in all age groups. Reliable and repeatable BP measurements in all age groups of children are best achieved with K5 as the indicator of diastolic BP.