The human papillomavirus has emerged over the past decade as the leading candidate to be the sexually transmitted aetiological factor in cervical cancer. Although it appears that papillomavirus types 16 and 18 are associated with a higher risk of advanced cervical neoplasia, most of the evidence comes from studies which do not satisfy basic epidemiological requirements, and are therefore difficult to interpret. The most significant problems are the small sample size, potentially biased selection of study subjects, the difficulties in cytologically distinguishing precancerous lesions from papilloma infection of the cervix, the unknown specificity and sensitivity of the various hybridisation methods for determining papillomavirus infection status, and the statistical analyses and presentation of results. On the basis of the existing studies, one is forced to conclude that, while experimental data suggest an oncogenic potential for HPV, the epidemiological evidence implicating it as a cause of cervical neoplasia is still rather limited.