Women with a family history of breast cancer generally self-refer because they have a feeling that their risk is high. However, they have, in general, only a hazy notion of the population risk of breast cancer and their own risk in relation to this. It is probable that they are helped by genetic counselling and, if at substantial risk, by annual mammography. However, the psychological impact of assigning true risk and the value of mammography need to be evaluated. We have assessed risk perception by questionnaire in 517 new referrals to a family history clinic and 200 women returning to the clinic at least 1 year after counselling. Correct assignment of population lifetime risk of breast cancer was 16% in the uninformed precounsel group and 33% in the post-counsel group, likewise personal risk was correct in 11% and 41% respectively. Post-counsel women were significantly more likely to retain information if they were sent a post-clinic letter or if they assessed their personal risk as too high initially.