National observational studies indicate that infants who are not restrained in child safety seats (CSSs) in motor vehicles are usually riding on the lap of another occupant. This study was undertaken to determine the conditions under which children travel on-lap. The extent to which injuries would be reduced if these children were restrained in CSSs was also examined. Data were taken from a multihospital monitoring system for pediatric occupant injuries and from the coroner's office in a single urban county (1980 through 1989). One hundred ten children younger than 1 year of age evaluated in the monitored emergency departments after involvement in a crash had been traveling on-lap. On-lap travel did not appear to result from overcrowding. Eighty-eight percent were riding in vehicles with five or fewer occupants; 83% were in the front seat; 58% sustained injury; 22% of those evaluated were hospitalized; and 15% sustained intracranial injury. A 30% reduction in overall injury, a 75% reduction in hospitalization, and a 69% reduction in intracranial injury were projected for those riding on-lap, had they been restrained in CSSs. Applying national rates of on-lap travel (16.8%) to National Highway Traffic Safety Administration estimates of the number of infants injured in crashes each year indicates that approximately 2218 of these children would be on the lap of another passenger. Substantial savings in terms of injury and associated costs can be realized if children traveling on-lap were in CSSs. Parent education as well as strict enforcement of CSS laws must be implemented.(ABSTRACT TRUNCATED AT 250 WORDS)