Appointments were made with parents to interview children in their homes. The study hypothesized that sibling relationship variables would condition the behavioral resemblance of younger and older siblings. For both delinquency and substance use, this prediction was confirmed for warmth and mutual friends. Sibling pairs who reported warmer mutual relationships or greater contact with mutual friends were more alike behaviorally. Statistical sibling effects were not explained by social class, parental substance use, or rearing styles. These effects were interpreted as the influence of one sibling on the other and as the influence arising from sharing common friends. The authors conclude that, given the existence of sibling effects, the strength of shared familial influences of other origins should be revised downward. 17 references and 4 tables