A group of 100 children who had experienced one episode of accidental poisoning were studied. A group of matched controls were selected from a previous randomized population derived from the City of Syracuse. The composition of study and control groups was similar in all gross characteristics tested. The poisoned group was found to have many more subsequent episodes of poisoning than the control group. These children were more active and mischievous, were more likely to experience falls, had more frequent separation experiences from the parents, and the maternal-child relationships appeared to differ significantly in several characteristics. It would appear that the repeater concept in accidental poisoning is established and presents a challenge to those presently concerned with prevention of accidental poisoning in children.