For more than half a century, numerous proposals have been advanced for the mode of action of carcinogens. This review presents a wide array of evidence that implicates oxidative stress (OS) in many aspects of oncology, including: formation of reactive oxygen species (ROS) by the major classes of carcinogens (as well as minor ones), cancer stages, oncogene activation, aging, genetic and infectious illnesses, nutrition, and the role of antioxidants (AOs). Although diverse origins pertain, including both endogenous and exogenous agents, ROS are frequently generated by redox cycling via electron transfer (ET) groups, e.g., quinones (or phenolic precursors), metal complexes (or complexors), aromatic nitro compounds (or reduced products), and conjugated imines (or iminium species). We believe it is not coincidental that these functionalities are often found in carcinogens or their metabolites. The pervasive aspects of DNA binding by ultimate carcinogens, and mutations caused by ROS are treated. Often, ROS are implicated in more conventional rationales, such as oncogenes. A multi-faceted approach to mechanisms appears to be the most logical. The OS unifying theme represents an approach which is able to rationalize the diverse data associated with carcinogenesis. Because this theoretical framework aids in the understanding of cancer initiation, it can serve as a useful tool in combating cancer, particularly in relation to prevention. Significantly, the electron transfer--oxidative stress (ET-OS) scenario can also be applied to many drug categories, toxins, enzymes, and hormones.