Professor Lev's paper is both interesting and suggestive, and I am glad to have the occasion to review it. After one or two remarks of a perspective nature, I should like to suggest an application of the method that I believe may be useful. It seems to me that, as happens to so many of us, Lev has got an inspiration from information theory, and then has gone ahead and used something else. His method of final adjustment of his cell entries does not adjust the information ratio, because the required algorithm for calculation has not been developed. Instead, he resorts to something like a maximum likelihood principle (the results of which are probably not much different from a minimization of the departures of the information content from the desired content) and then, as another approximation, uses Stephan's method for adjusting the cell entries until he has converged to within some reasonable distance of the least-squares solution to the maximum likelihood entries. Personally, I see nothing wrong with this procedure, but it does leave room for some further theoretical research, or perhaps some experimental calculations designed to test the goodness of fit of the results obtained in this way. If the method were not so simple, so eminently plausible, it might be desirable to go back to the original business situations in which the several forecasts were required, and to ask what degree of accuracy was required. In my opinion we don't need to do this, because Lev's method makes sense, is simple in application, and does at least as well as any other plausible method. There is a useful application in the field of planning. Often we have to plan a two-way distribution of effort-for instance, by project by time period, or by manhours by project, etc. We find that we can plan the