Childrens' touching and mouthing behaviors during outdoor play in urban residential yards were measured using video observations. Descriptions were made of childrens' outdoor residential play environments. Behaviors assessed were used to examine (1) validity of parental responses to questions on childrens' oral behaviors and outdoor play and (2) relationships of mouthing behaviors to blood lead levels (BLLs). Thirty-seven children aged 1-5 years were recruited for 2 h of video recording in their yard and blood lead measurement. Video assessments included hourly rates of hand touches to ground/walking-level surfaces (cement/stone/steel, porch floor/steps, grass, and bare soil) and oral behaviors. Parental questionnaires assessed their child's outdoor activities, behaviors, and home environment. The children were: mean 39 months; 51% male; 89% Hispanic; and 78% Medicaid or uninsured. Twenty-two children had a blood lead measured (mean 6 microg/dl). During taping, all children had access to cement, 92% to grass, 73% to bare soil, and 59% to an open porch. Children had frequent touching and mouthing behaviors observed (median touches/h: touches to surfaces 81; hand-to-mouth area (with and without food) 26; hand-in-mouth 7; and object-in-mouth 17). Blood lead was directly correlated with log-transformed rates of hand-in-mouth (Pearson's correlation, r=0.564, n=22, P=0.006) and object-in-mouth (Pearson's correlation, r=0.482, n=22, P=0.023) behaviors. Parental questionnaire responses did not accurately reflect childrens' observed oral behaviors, play habits, or play environment. These data confirm the direct relationship between hand-to-mouth activities and BLLs and fail to validate parental perceptions of their child's mouthing behaviors or outdoor play environment.