The sample consisted of children under age 18 who were evaluated for possible sexual abuse from 1987 to 1989. The demographic, historical, and physical findings of the cases were analyzed. T-test, analysis of variance, and chi-square analyses were used to test for differences related to ethnicity, race, gender, and age. Significant differences in these factors were found with respect to the relationship of the perpetrator to the child, the number of perpetrators, the interval from the abuse to the evaluation, the types of abuse, and the findings of the physical examination. In 77 percent of the cases, the alleged perpetrator was someone the child knew. Many of these cases involved preverbal children suspected of being abused by a parent. Mexican-American children were more likely than Anglo or African-American children to be abused by an extended family member. Nine percent of the children reported abuse episodes that involved more than one perpetrator. African-American children were much more likely to be abused by more than one person. The mean time from the last abusive episode to the time of the evaluation was 65 days. Mexican-American children presented after a longer interval than Anglo or African-American children. This ethnic difference was reduced once the child's sex and the relationship of the perpetrator to the child were also considered. In 48 percent of cases, there was a clear history of penetration. African-American children reported much higher rates of penetration than Mexican-American or Anglo children. Differences in patterns of abuse were also found by gender and age. The authors advise that this study of the complex influences of ethnicity, race, gender, and age on characteristics of child sexual abuse contributes to a better understanding of contributing factors, the potential impact of the abuse on the child and the family, legal implications, and preventive measures. 3 tables and 28 references