Goals were to determine if a sufficient number of adults had enough flexible time and emotional resources to take on the demands of mentoring at-risk youth, whether mentoring could be integrated into large-scale youth-serving institutions, features of the adult role in effective mentoring relationships, the level of training and support activities required to effectively administer mentoring programs, and whether participants in mentoring programs made observable changes in behaviors and attitudes. The first product of the research initiative, an evaluation of four Linking Lifetimes programs, identified and defined effective adult/youth relationships. Interviews were conducted with program participants in Massachusetts, Tennessee, California, and Florida. At each site, elders were required to meet at least weekly with youth for 4 to 10 hours. Elders received stipends ranging from $2.20 to $6.00 per hour and reimbursement for expenses. Of 26 adult/youth pairs, 17 were satisfied with the relationship and 9 were dissatisfied. Particular activities engaged in by pairs did not determine satisfaction. Differences were observed, however, in participant interaction styles. Elders in satisfying relationships allowed the relationships to be youth-driven in content and timing. Mentors in unsatisfying relationships were more likely to follow the ineffective pattern of thinking youth had no activity preferences. Other distinguishing features between satisfied and unsatisfied pairs concerned the willingness of young people to discuss intimate details about their families and themselves, how mentors offered support and advice to youth, adult attitudes toward youth, and adult involvement with families. Program practices that promote the effectiveness of mentoring are considered, and recommendations to enhance mentoring programs are offered. Supplemental information on the Linking Lifetimes program is appended. 33 references and 7 tables