To the Editor:—
Dr. Philip Rubin can be commended in general for his presentation of the present knowledge and controversy in the study and management of breast cancer (199:732, 1967). However, he has made an error in interpretation of Fig 1, which he states as illustrating that "the age distribution of total cases of breast cancer is at variance with the annual incidence of all cancer, skewing the curve into a younger population."In the first place, these two curves are at variance by definition, since one is a relative frequency distribution of patients' ages and the other is a line graph of age-specific inci-dence rates; the two curves are not comparable under any circumstances. Secondly, by referring back to Dr. Rubin's source, one discovers that he has borrowed Berkson's poorly labelled graph without the benefit of the statistician's accompanying text which clarifies and correctly describes the figure.