To determine the relationship between activity levels of parents and those of their young children, we monitored physical activity with a mechanical device, the Caltrac accelerometer, in one hundred 4- to 7-year-old children and in 99 of their mothers and 92 of their fathers. During 1 year in the Framingham Children's Study, data were obtained for an average of more than 10 hours per day for 8.6 +/- 1.8 days for the children, for 8.3 +/- 2.1 days for their mothers, and for 7.7 +/- 2.3 days for their fathers. Children of active mothers (average Caltrac accelerometer counts per hour greater than the median) were 2.0 times as likely to be active as children of inactive mothers (95% confidence interval = 0.9, 4.5); the relative odds ratio of being active for the children of active fathers was 3.5 (95% confidence interval = 1.5, 8.3). When both parents were active, the children were 5.8 times as likely to be active (95% confidence interval = 1.9, 17.4) as children of two inactive parents. Possible mechanisms for the relationship between parents' and child's activity levels include the parents' serving as role models, sharing of activities by family members, enhancement and support by active parents of their child's participation in physical activity, and genetically transmitted factors that predispose the child to increased levels of physical activity.