The etiology of chronic fatigue syndrome (CFS) has been both obscure and highly contentious, leading to substantial barriers to both clear diagnosis and effective treatment. I propose here a novel hypothesis of CFS in which either viral or bacterial infection induces one or more cytokines, IL-1beta IL-6, TNF-alpha and IFN-gamma. These induce nitric oxide synthase (iNOS), leading to increased nitric oxide levels. Nitric oxide, in turn, reacts with superoxide radical to generate the potent oxidant peroxynitrite. Multiple amplification and positive feedback mechanisms are proposed by which once peroxynitrite levels are elevated, they tend to be sustained at a high level. This proposed mechanism may lower the HPA axis activity and be maintained by consequent lowered glucocorticoid levels. Similarities are discussed among CFS and autoimmune and other diseases previously shown to be associated with elevated peroxynitrite. Multiple pharmacological approaches to the treatment of CFS are suggested by this hypothesis.