The cause of most cancers is unknown, appearing apparently sporadically in the population. Where causality is known, some have a single cause only. It is hypothesised that some cancers, currently of unknown cause, are due to a single cause. Databases relating to the aetiology and epidemiology of a number of selected cancers and to theories of cancer causation were systematically searched. Data review shows that some cancers have a single cause, even in all known settings; others have single causes in specific settings and different causes in alternative settings; others have multiple causes. It is further hypothesised that, particularly for breast and ovarian cancer, specific directed investigations might be successful in elucidating the aetiology if due to single cause. Such investigations should occur in groups of diseased subjects, non-diseased but at-risk subjects, and normal (control) populations. These investigations should include tests for trace elements, chemicals, antibodies to infective agents, and markers of inflammation. Systematic differences between groups might give strong clues to aetiology.