OBJECTIVE--To identify the nature of pain and discomfort experienced during mammography and how it can be ameliorated. DESIGN--Questionnaire survey before invitation for mammography and immediately after mammography. Responses before screening were related to experience of discomfort. SETTING--Health district in South East Thames region. SUBJECTS--1160 women aged 50-64 invited routinely for screening; 774 completed first questionnaire, of whom 617 had mammography. 597 completed the second questionnaire. MAIN OUTCOME MEASURES--Reported discomfort and pain, comparisons of discomfort with that experienced during other medical procedures, qualitative description of pain with adjective checklist. RESULTS--35% (206/597) of the women reported discomfort and 6% (37/595) pain. 10 minutes after mammography these figures were 4% (24/595) and 0.7% (4/595) respectively. More than two thirds of women ranked having a tooth drilled, having a smear test, and giving blood as more uncomfortable than mammography. The most important predictor of discomfort was previous expectation of pain (discomfort was reported by 21/32 (66%) women who expected pain and 186/531 (35%) who did not). Discomfort had little effect on satisfaction or intention to reattend. CONCLUSIONS--The low levels of reported pain and discomfort shortly after mammography and the favourable comparisons with other investigations suggest that current procedures are acceptable. Since two thirds of the women experienced less pain than expected health education and promotion must ensure that accurate information is made available and publicized.