A retrospective study was designed to test J.N. Wolfe's theory that xeromammographic breast patterns can be used as a risk index for breast cancer. After analyzing breast types of all breast cancer patients over a 4 year period at the Jewish General Hospital, Montreal, who had xeromammography prior to surgery, we were unable to support Wolfe's hypothesis. We conclude that an inherent bias in the design of his experiment led to erroneous conclusions.