In 1982, 113 first- and second-graders attending an elementary school in West Jerusalem had their blood pressure (BP), height, weight, and resting pulse measured. Three successive BP readings were taken. The mean of the second and third readings was used for data analysis. Mean systolic blood pressure (SBP) was 98.30 +/- 8.51 mm Hg and mean diastolic blood pressure (DBP) was 67.63 +/- 6.85 mm Hg. An analysis of covariance was performed in order to determine the contribution of each independent variable (height, weight-for-height, sex, age, and resting pulse) to the variance in BP, while adjusting for all other independent variables. Height, weight-for-height, and resting pulse were all found to be significantly associated with both SBP and DBP. Children in the upper quartile of weight-for-height had a mean adjusted SBP that was 10.2 mm Hg higher than those in the lower quartile and a mean adjusted DBP that was 7.9 mm Hg higher. Children in the upper quartile of height (for age) had mean adjusted SBP and DBP that were 6.1 and 4.7 mm Hg higher, respectively, than those of children in the lower quartile. Sex and age showed no significant association with SBP. The associations between DBP and both sex and age were not statistically significant (P less than 0.10), but there was a trend for girls and older children to have a slightly higher DBP. In a subsample of 93 children who had both resting pulse and recovery index measured (using the Modified Harvard Step Test for children).(ABSTRACT TRUNCATED AT 250 WORDS)