The use of self-report measures for the assessment of depression in children has gained wide popularity as a component in the evaluation of children's mental health and well-being. However, the extent to which children are reliable and consistent reporters of their depressive symptoms is sometimes called into question. This study examined the test-retest reliability of children's reports of depressive symptomatology on the Reynolds Child Depression Scale (RCDS; Reynolds, 1989a). The sample consisted of 220 children, in grades 3 through 6, representing a cross-section of ethnic groups. Children were tested twice, with a 4-week interval between testings. Results showed a high degree of stability in children's responses to the RCDS, with a test-retest reliability coefficient of .85 for the total sample, and a mean difference between testings of less than 2 points. Test-retest reliability coefficients were also computed for males and females, for each grade, and for white, black, Hispanic, and Asian children. Overall, the results lend strong support for the reliability and stability of children's self-reported depressive symptomatology as assessed by the RCDS.