Regular, vigorous physical activity (aerobic exercise) appears to have significant benefits in preventing disease, but exercise levels continue to be low in spite of the wide availability of intensive fitness programs. Self-instructional behavior-change packages can reach more people than face-to-face methods, and can address a range of problem behaviors. This study investigated the effectiveness of a self-instructional training program for aerobic exercise. Participants were allocated randomly to an exercise correspondence course involving several mailings of information (n=53), or to exactly the same program mailed in a single package (n=52). People who initially showed interest in the course but withdrew before it began (n=33), and participants in standard fitness classes (n=31), were used as comparison groups. All courses were 12-week aerobic programs of gradually increasing intensity. At the end of the program, participants in the single-package course were significantly more active than those in the multiple-mailing program, and were similar to those in the fitness class. At a 10-month follow-up, there were no significant differences among the reported physical activity levels of participants in the four different conditions. Despite the somewhat weak effects obtained in this investigation, programs that can be administered by mail can reach large numbers of people who may wish to change health-related behaviors.