The sequential mammograms of 48 patients (median age: 54 years, range: 24-84 years) undergoing primary medical treatment for large operable breast cancer at this institution were reviewed, and compared to the findings at clinical assessment. Twenty-six patients underwent endocrine therapy using tamoxifen, and 22 chemotherapy (CMF or MMM). All had more than two mammograms which were reviewed blindly by two independent radiologists. Response to treatment by both evaluation modalities was recorded using WHO definitions (NC, PR, CR, PD). Some 81% of patients achieved an objective clinical response to treatment, whilst 63% showed mammographic response. Overall comparison of clinical examination with mammography showed agreement in 38 patients (79%) and disagreement in 10 (21%). Agreement in type of response but not in its degree was found in 22 cases (46%). The results suggest that in the majority of cases mammography provides a useful adjunct to clinical examination in monitoring response to primary medical treatment for breast cancer. However, radiological factors such as unchanging microcalcification, and the continuing presence of mammographic density contribute to the discrepancy with clinical response in some tumours, and the search for alternative modalities of assessment should continue.