Research has revealed that exercise is effective for reducing symptoms of depression and anxiety. The mechanisms by which these reductions occur, however, have not been widely studied. To examine several potential theories, a prospective, randomized, 7-week exercise intervention was conducted. Untrained participants were randomly assigned to an aerobic exercise group or to a stretching-control group. Participants completed several questionnaires to assess psychological variables, including measures of depression and anxiety, and blood was drawn at pre- and post-test to measure serum serotonin levels. A mixed-design ANOVA revealed that the exercise group had lower levels of depression than the stretching-control group after the intervention. The exercise group also showed a larger percentage decrease in serotonin than the stretching-control group. This reduction in blood serotonin after exercise is similar to the effects of selective serotonin reuptake inhibitors. Additionally, percent change in serotonin was found to partially mediate the relationship between exercise and depression.