Whether a drug such as levodopa, which is prescribed for long periods, may be toxic is a legitimate and even indispensable question. The problem is no different from that posed by other drugs-such as calcium antagonists, antihypertensives, or hormones-normally prescribed for chronic diseases. What, however, is meant in this context by "toxic" (from the Greek toxicon, meaning poison)? Irrevocable damage such as cell loss should not be confused with reversible side effects resulting from cell dysfunction. Clinically or experimentally, levodopa has not been shown to accelerate neurodegeneration or cause permanent impairment of cell function in a manner that would result in irreversible side effects. These data have been reasonably well established in vivo in animals and humans, although preliminary studies suggesting that levodopa is a trophic factor remain unconfirmed. Like oxygen or calcium, levodopa can be toxic in vitro when it is present in high concentrations or in the absence of glial cells. However, glial cells are much more numerous than neurons in vivo, so these conditions cannot simply be extrapolated to three-dimensional brain structures in which protective interactions with the cellular environment abound. Because levodopa remains the most effective treatment available for Parkinson's disease, questions regarding timing or manner of administration of the drug should arise not because levodopa is toxic to nerve cells, but because it causes reversible side effects. When the elementary rules of substitutive therapy to provide maximum comfort while limiting side effects are followed, we need not fear that levodopa is dangerous unless the contrary is proven.