Possible risk factors for cervical cancer were investigated in 645 women, 20-49 years of age from Copenhagen, with histologically confirmed cervical cancer or carcinoma in situ and in 614 controls drawn at random from the female population of the same area. To study the role of the "male factor", monogamous cases and controls together with their husbands were invited for further examination. This included a personal interview, a blood sample for analysis of herpes simplex virus antibodies, and a penile swab for papillomavirus analysis. In total, 41 case couples and 90 control couples were enrolled (89% and 92% of eligibles, respectively). The most significant risk determinants were a history of genital warts in the male (RR = 17.9 for ever vs. never) and ever having used condoms (RR = 0.2). Other potentially important factors, although not statistically significant, were ever having visited prostitutes (RR = 2.6, p = 0.14), circumcision (RR = 0.3, p = 0.18), and a history of genital warts in the female (RR = 4.6), p = 0.09). Having a husband with a history of multiple sexual partners was also associated with an increased crude risk, but when adjustment was made for husband ever having had genital warts, use of condom, and age, no significant excess risk was observed. Human papillomavirus DNA, which was tested for with ViraPap and ViraType, was detected in one of the controls, and in only 2 case husbands. This study points to the importance of "high-risk males" for the development of cervical cancer in their sexual partner; the results support the infectious nature of the disease.