Professor Melicher has conducted an interesting study of the effects of financial factors on beta coefficients and their variations. He reduced the perennial multicollinearity problem by using a factor analysis to screen the financial variables used in the statistical analysis. This study is another in the growing body of literature concerned with beta coefficients as measures of risk. My comments on this paper will consider separately (1) the research design (i.e., what the study directly covered), and (2) the inferences drawn by the author versus what the beta coefficients and their variation really do measure.