FROM time to time, groups of surgeons betake themselves to prove, by statistical means, that survival of patients with cancer of the breast following simple mastectomy is equal to or better than the survival time following radical mastectomy. I am certain the motivation of these groups to publish such statistical results stems from a scientific desire to measure and evaluate different procedures as they may affect the course of a disease. Many of us react to these statistical reports with the feeling that the principles of cancer surgery are being sacrificed for statistical illusions. Yet, it is difficult to incriminate the individual statistical data so offered. Let us examine a few typical series in hope of discovering a denominator common to, and responsible for, the clinical assumptions from which statistical analyses and evaluations have been made. In 1956, Hickey et al reported on a series of 1,661 patients with breast