An evaluation of the contribution of wheat breeding to increasing yields on farms can be accomplished by an examination of the effect on farm production of changes in varieties and other inputs using a production function. Three different means of representing the effect of varietal change in a production function are identified, and the advantages and limitations of these are discussed. An empirical analysis of each is presented for the Mitchell Shire in southern N.S.W. The evidence suggests that the different measures are not highly correlated, so the choice between them can be critical to the results obtained from a full examination in a production function.