TECHNOMETRICS

Dr. Cornell and Dr. Ott are to be commended for their interesting and useful paper on the interpretation of mixture response surfaces. My comments will be directed toward the usefulness of the techniques in practice, ways in which the methodology is more general than the authors imply, and situations where the results may not be representative. I will also point out some similarities between Cornell and Ott's method of response surface interpretation and the construction of screening designs for mixtures. The rocket propellant (Kurotori) and flare experiment (McLean and Anderson) data sets discussed by the authors will be used to illustrate these points.