Experimental and clinical data on mammography as a tool for population screening are reviewed. The conclusions are that [1] there is good evidence that annual history, physical examination, and mammography can reduce short-term and midrange breast cancer mortality by about one third; [2] the evidence that mammography alone plays a significant role in this reduction is weak and indirect; [4] data on long-term effects of mammography are lacking; [4] no satisfactory investigations of associated radiation hazards have been published; and [5] the possible benefits of mammography have received more emphasis in the clinical literature than have its defects. Promotion of mammography as a general public health measure is premature.