Children as young as age 8 reported sexual activity (defined as having engaged in sexual intercourse), and more than one-third of the children age 8 to 18 reported being sexually active. Of those who were sexually active, more than one-third were not using contraceptives. Using logistic regression, five variables are identified as having importance in explaining sexual activity: age, race, behavior problems, homicidal ideation and drug use. Two variables have some limited ability to explain contraceptive use: gender and entering care because of parental absence or incapacity. Findings suggest that many children in care who are sexually active are very troubled and need multidimensional assessment and treatment, interventions that encompass the constellation of problems rather than each in isolation. The article discusses implications of these and other findings and makes suggestions for policy and practice. Tables, references