The cause of Parkinson's disease (PD) is unknown, but reduced activity of complex I of the electron-transport chain has been implicated in the pathogenesis of both mitochondrial permeability transition pore-induced Parkinsonism and idiopathic PD. We developed a novel model of PD in which chronic, systemic infusion of rotenone, a complex-I inhibitor, selectively kills dopaminergic nerve terminals and causes retrograde degeneration of substantia nigra neurons over a period of months. The distribution of dopaminergic pathology replicates that seen in PD, and the slow time course of neurodegeneration mimics PD more accurately than current models. Our model should enhance our understanding of neurodegeneration in PD. Metabolic impairment depletes ATP, depresses Na+/K(+)-ATPase activity, and causes graded neuronal depolarization. This relieves the voltage-dependent Mg2+ block of the N-methyl-D-aspartate (NMDA) subtype of the glutamate receptor, which is highly permeable to Ca2+. Consequently, innocuous levels of glutamate become lethal via secondary excitotoxicity. Mitochondrial impairment also disrupts cellular Ca2+ homoeostasis. Moreover, the facilitation of NMDA-receptor function leads to further mitochondrial dysfunction. To a large part, this occurs because Ca2+ entering neurons through NMDA receptors has 'privileged' access to mitochondria, where it causes free-radical production and mitochondrial depolarization. Thus there may be a feed-forward cycle wherein mitochondrial dysfunction causes NMDA-receptor activation, which leads to further mitochondrial impairment. In this scenario, NMDA-receptor antagonists may be neuroprotective.