To address the hypothesis that male sexual behavior may affect the etiology of invasive cervical cancer, a case-control study was undertaken in Panama, Costa Rica, Colombia and Mexico. The study enrolled husbands of those women with invasive cervical cancer and of those age-matched controls who reported only one lifetime sexual partner. A total of 204 case and 485 control husbands (78% and 72%, respectively, of identified husbands) were interviewed, clinically examined, and had penile swabs taken for papillomavirus assays. Risk increased significantly (p = 0.005) with the number of sexual partners reported by the husband (RR = 2.0 for 26+ vs. <6 partners). Low educational status of the husband was also an important predictor of risk, possibly indicating the role of unmeasured aspects of sexual behavior. Visits to prostitutes, circumcision status and sexually transmitted disease histories were not important predictors of risk, but evidence from clinical examination indicated that poor genital hygiene may be involved. Human papillomavirus (HPV) expression as defined by filter in situ hybridization was detected in 20-23% of subjects and, except in the small group with both HPV types 6/11 and 16/18, was not related to risk. This may reflect sampling problems in the male or the importance of host factors which enhance viral carcinogenicity in the female.