A woman's risk of cervical cancer is generally thought to be related to her sexual behaviour. The sexual background of her male partners is also important. In some societies, a woman's risk of cancer of the cervix will depend less on her own behaviour than on that of her partner. Male sexual behaviour, particularly in relation to prostitution, may account for two hitherto unexplained features of the epidemiology of this disease--the extremely high incidence in Latin America and the decline in mortality this century. If this is so and men carry the aetiological agent, it will be important to discover whether they do so for short or long periods.