Objective: To evaluate galactography and cytology in women with nipple discharge without clinical or mammographic evidence of cancer. Methods: During a 12.5-year period, 384 women (15–85 years, mean age 47.5 ± 14 years) were referred for galactography and smear cytology for recent onset of spontaneous, non-milky nipple discharge. Patients with clinical or mammographic evidence of tumor underwent excisional biopsy directly. Among 314 galactograms, 189 [60.2%; 95% confidence interval (CI) 54.5%, 65.6%] biopsies were recommended. A further 11 patients were scheduled for biopsy because of mammography or cytology. Results: Sixteen of 182 biopsied patients had malignancies (8.8%; CI 5.3%, 14.1%). Combined rate of papillomas, papillomatous proliferation, and malignant tumors was 59.9% (109 of 182; CI 52.4%, 67.0%). Biopsy was malignant in three of 56 women (5%) with nonhemorrhagic discharge and in 13 of 97 (13%) with hemorrhagic discharge (P = .26). Exfoliative cytology revealed 11 false-negatives, four false-positives, five true-positives, and 153 true-negatives (sensitivity 31.2%, CI 11%, 58%; specificity 97.4%, CI 93%, 99%). In ten of 158 patients (6.3%) with suspicious galactography, cancer was found by biopsy. Sensitivity of galactography for malignancy was 83% (CI 51.6%, 97.9%) and specificity was 41% (CI 35.2%, 46.5%). Galactographic sensitivity for any (benign or malignant) neoplasm was 94% (93 of 99; CI 87%, 98%) and specificity was 55% (119 of 215; CI 48%, 62%). Half of the cancers were detected exclusively by galactography. Conclusion: Cytology is helpful when positive and galactography localizes the source of discharge. Biopsy is indicated when palpation, mammography, cytology, or galactography is suspicious.