Selegiline monotherapy has been clearly demonstrated to delay the development of disability in early, otherwise untreated Parkinson's disease patients. It remains uncertain, however, whether this benefit is due to protective effects on residual neurons or to symptomatic effects that mask the detection of underlying disability. This paper examines the evidence and theory supporting the hypotheses that selegiline acts by protective or symptomatic mechanisms and considers what future studies might help clarify these issues.