Analogous to the epidemiology of different types of human papillomaviruses (HPVs), variants of a single type such as HPV16 may differ in the natural course of infection. A prospective study was conducted in a cohort of female university students who did not have HPV16 DNA or antibody at enrollment. Subjects were followed every 4 months with Pap smear and colposcopic examinations and tests for HPV DNA and antibody. Of 528 women, 62 acquired HPV16 infection during follow-up. The 5-year cumulative incidence was 12.6% for infection with HPV16 prototype-like variants and 3.1% for infection with non-prototype-like (NPL) variants. Among those with incident HPV16 infection, the adjusted odds ratio of detecting NPL variants was 6.0 [95% confidence interval (CI), 1.3-27.5)] for nonwhite women or white women who had had nonwhite sex partners compared with white women who had not had nonwhite sex partners. As compared with women who had not used hormonal contraceptives in the 8 months prior to the first HPV16-positive detection, the adjusted odds ratio of detecting NPL variants increased from 1.6 (95% CI, 0.3-9.1) for those who had used hormonal contraceptives for 1-4 months to 5.4 (95% CI, 1.0-28.3) for those who had used hormonal contraceptives for 5-8 months (P for trend = 0.04). There was no difference in HPV16 seroconversion rates between women with prototype-like variants and women with NPL variants. The increased risk for biopsy-confirmed cervical intraepithelial neoplasia grades 2 or 3 associated with detection of incident infection by NPL variants was not explained by differences in persistence between the two variant groups, indicating that biological mechanisms other than viral persistence may be responsible for the observed difference in risk for cervical intraepithelial neoplasia grades 2 or 3.